avg"?!
.«dufil‘o...

Wag ....

.5".ch

14?»). r. .
§$§.
(my hr“:-

b.

y .1 . V ‘ ‘ . . . . \
muggy? at.-. ‘ . . , é.

 

llllllllllllllllllllllll(Hill’ll’lllllllllllllll

293 01019 0837

This is to certify that the

thesis entitled

Risk Analysis of Subirrigation Investment
Decisions in the Saginaw Bay Area
of Michigan

presented by

Katherine Kampmann

has been accepted towards fulfillment
of the requirements for

Mo S 0 degree in AgfiCUltural

 

Economics

   

{Ed/1:”

Major professor

Date 18 February 1993

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

“—1 7 , —._ . ~ 7 7, 7 -—* 7 ,,,i

 

LIBRARY
Michigan State
University

 

 

 

PLACE N RETURN BOXtoromovothb chockoutfrom yomrooord.
TO AVOID FINES return on or More data duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU ioAn Affirmative ActioNEqmi Opportunity lmtitwon

 

 

m1

 

RISK ANALYSIS OF SUBIRRIGATION INVESTMENT DECISIONS
IN THE SAGINAW BAY AREA OF MICHIGAN

BY

KATHERINE KAMPMANN

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Agricultural Economics

1993

ABSTRACT
RISK ANALYSIS OF SUBIRRIGATION INVESTMENT DECISIONS
FOR CORN PRODUCTION IN THE SAGINAW BAY AREA OF MICHIGAN
By

Katherine Kampmann

This thesis examines on-farm investment and operating costs and financial
benefits of two improved drainage and Six subirrigation investments on a representative
field in the Saginaw Bay Area of Michigan. Net present values (NPVS) are generated
from a base sequence and from random sequences of corn yields Simulated from
historical weather data in Michigan. Under both certainty and risk, the surface water
subirrigation system at 60-ft tile Spacing is the most profitable investment. For the 1958-
87 actual weather sequence, the existing drainage-only system is dominated by two
surface and one well water source subirrigation systems and dominates the other
subirrigation and narrower-spaced drainage-only systems. Under random weather
conditions, subirrigation with surface water at 30- and 60-ft tile spacings dominates the
other systems by first or second degree stochastic dominance, so long as the price of

corn remains above $2.05/bu (or $3.00/bu for the well water source).

This thesis is dedicated to Jeff, who saw me through the endeavor, and my parents, who

have always encouraged me to endeavor.

ACKNOWLEDGMENTS

I could not have completed this thesis without the help of a host of collaborators.

My major advisor, Dr. Scott Swinton, provided the financial support that kept the
bills paid, but more importantly with the academic support that helped shape the
direction and outcome of the research. He devoted many hours advising me and making
extensive comments and suggestions concerning the research and the drafts of the thesis.
I feel lucky to have had such a devoted thesis advisor.

Dr. Harold Belcher of the Department of Agricultural Engineering, the principal
investigator for the Water Quality Impacts of Water Table Management Project,
provided guidance concerning the technical aspects of subirrigation and the simulation
model, DRAINMOD.

Dr. Roy Black of the Department of Agricultural Economics contributed helpful
comments and advice concerning the theoretical framework of the economic analysis.

James LeCureux, an extension agent in Huron County, contributed greatly to this
thesis through his own research on subirrigation and by providing Huron County Specific
data used in the analysis. He was always available to answer the many questions I had in
the initial stages of the research.

Without the computer support of Brian Baer and without the soil input data
provided by Martin Rosek, I could have never gotten the DRAINMOD simulations to

run.

iv

I would also like to thank all of the many other people who provided the input
data necessary to run the simulations and perform the economic analysis. These people
include Dr. Fred Numberger and Dr. Jeffrey Andresen of the Department of
Agriculture, Environmental Division, Department of Climatology; Jim Angel of the
Midwest Regional Climate Center in Champaign, Illinois; Chip Cheschire of the
Department of Biological and Agricultural Engineering at North Carolina State
University; Dr. Chensheng He of the Department of Resource Development; Dr.
Francis Pierce and Dr. Maurice Vitosh of the Department of Crop and Soil Sciences;
the subirrigation contractors who provided cost estimates and who must remain
anonymous; Neil Krieger of Michigan Valley Irrigation; and Don Long and Jim
Murdock, two farmers using subirrigation in Huron County.

I would also like to thank Dr. John Staatz of the Department of Agricultural
Economics, who provide the financial support that helped me make it through the first
two years of my Master’s program.

Finally, this research was made possible through funding from the Michigan
Agricultural Experiment Station under the special grant: Water Quality Impacts of Water

Table Management.

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
CHAPTER 1: INTRODUCTION
1.1. Introduction
1.2. Objectives
1.3. Methods
CHAPTER 2: LITERATURE REVIEW
2.1. Introduction
2.2. Technical Aspects of Subirrigation
2.2.1. Subirrigation Defined and Described
2.2.2. Soil and Land Characteristics
2.2.3. Water Availability
2.2.4. Water Quality Aspects of Subirrigation
2.3. Economic Analyses of Irrigation

2.4. Economic Evaluations and Studies of Yield with
Subirrigation

2.4.1. Field Studies
2.4.1.1. Yield Analyses
2.4.1.2. Economic Analyses
2.4.2. Simulation Studies
2.4.2.1. Benefit to Drainage
2.4.2.2. Benefit to Drainage/Subirrigation

2.5. Directions for the Current Economic Analysis of
Subirrigation

12
14
17

20

25
26
26
29
45
48

53

2.6. Gaps in the Study
CHAPTER 3: METHODOLOGY
3.1. Introduction

3.2. Simulation

3.2.1. Simulation as a Tool in Economic Analyses
3.2.2. DRAINMOD
3.2.2.1. Model Validation

3.3. Economic Analysis

3.3.1. Base Calculations
3.3.2. Net Present Value Analysis
3.3.2.1. Theory of Profit Maximization
3.3.2.2. Procedures
3.3.3. Monte Carlo Simulation
3.3.4. Risk Analysis
3.3.4.1. Risk Efficiency Models
3.3.4.2. Application of Stochastic Dominance Criteria
to Water Table Management Investment Decisions
3.3.4.3. Application of Expected Net Present Value Criteria
to Water Table Management System Investment Decisions

3.4. Sensitivity Analysis
CHAPTER 4: RESULTS
4.1. Simulation Results

4.1.1. Drainage Only Results
4.1.2. Subirrigation Results

4.2. Results of the Economic Analysis

4.2.1. NPV Analysis - Base Weather Sequence

4.2.2. Net Present Value and Expected Value-Variance

Analysis

4.2.3. Stochastic Dominance Analysis

4.2.4. Stochastic Dominance with Respect to a Function Analysis

4.3. Sensitivity Analysis
4.3.1. Sensitivity to Potential Yield

4.3.2. Financial Parameter Sensitivity Analysis
4.3.3. 4.3.3. Investment Cost and Output Price Sensitivity Analysis

“Ci

54
56
56
59
59
60
66
73
73
82
82
82
84
86
86
94
96
96
98
98

98
100

103
103
111
117
119
120
120

122
123

CHAPTER 5: SUMMARY AND CONCLUSIONS
5.1. Summary

5.2. Conclusions

APPENDICES

APPENDIX A: BASIC CODE FOR MONTE CARLO SIMULATION ENPV
AND SDNPV CALCULATIONS FOR WTMS ANALYSIS

APPENDIX B: DRAINMOD DATA INPUTS
APPENDIX C: DRAINAGE ONLY SITE DESIGN
APPENDIX D: SUBIRRIGATION SITE DESIGN

APPENDIX E: AVERAGE MONTHLY TEMPERATURE AND
PRECIPITATION FOR BAD AXE, HARBOR BEACH,
AND FLINT, MICHIGAN FOR THE PERIOD 1951-1980

APPENDIX F: DRAINMOD WATER BALANCE VERIFICATION
APPENDIX G: INVESTMENT COSTS
APPENDIX H: SIMULATION YIELD RESULTS

H1: DR20 - Drainage Only at 20-Ft Tile Spacings
H2: DR30 - Drainage Only at 30-Ft Tile Spacings
H3: DR60 - Drainage Only at 60-Ft Tile Spacings
H4: 8120 - Subirrigation at 20-Ft Tile Spacings
H5: 8130 - Subirrigation at 30-Ft Tile Spacings
H6: SI60 - Subirrigation at 60-Ft Tile Spacings

APPENDIX 1: SIMULATION WATER BALANCE RESULTS

Il: DR20 - Drainage Only at 20-Ft Tile Spacings
I2: DR30 - Drainage Only at 30-Ft Tile Spacings
13: DR60 - Drainage Only at 60-Ft Tile Spacings
I4: 8120 - Subirrigation at 20-Ft Tile Spacings
15: $130 - Subirrigation at 30-Ft Tile Spacings
I6: SI60 - Subirrigation at 60-Ft Tile Spacings

BIBLIOGRAPHY

127
127
132

135

135
143
151

152

153
154
160
163

163
164
165
166
167
168

169

169
170
171
172
173
174

175

TABLE 2.1:

TABLE 2.2:

TABLE 2.3A:

TABLE 2.3B:

TABLE 2.4:
TABLE 2.5:
TABLE 2.6:
TABLE 2.7:
TABLE 2.8:
TABLE 3.1:

TABLE 3.2:

TABLE 3.3:

TABLE 3.4:
TABLE 3.5:

TABLE 3.6:

TABLE 3.7:
TABLE 3.8:

TABLE 3.9:

LIST OF TABLES

Summary of Yield Results for 1987

Summary of Field Trials and Economic Analyses for 1988
Benefit to Splitting Tiles for Sugar Beet - 81 Treatment
Benefit to Splitting Tiles for Sugar Beet - DR Treatment
Summary of Field Trials and Economic Analyses for 1989
Summary of Field Trials and Economic Analyses for 1990
Yield and Net Return Results with Fair Surface Drainage
Results for Corn for a Rains Soil: Fair Surface Drainage
Results for Soybeans for a Rains Soil: Fair Surface Drainage
Summary of DRAINMOD Inputs

Genesee County Historic and Detrended Yield Compared with
DM Yields

Reese Farm Historic and Detrended Yield Compared with
DM Yields

Description of Investment Options
Summary of Component Costs for a WTMS

Total Investment and Annualized Per-Acre Investment Cost for
a 40-Acre System

Variable Costs Associated with Water Management Systems
System Repair and Maintenance Costs

Irrigated and Nonirrigated Corn Production Costs

31

34

35

36

40

44

50

52

52

63

70

71

75

77

78

8O

80

81

TABLE 3.10:

TABLE 4.1:

TABLE 4.2:

TABLE 4.3:
TABLE 4.4:
TABLE 4.5:
TABLE 4.6:
TABLE 4.7:
TABLE 4.8:
TABLE 4.9:

TABLE 4.10:

Base Parameter Values

DRAINMOD Yield Output for Drainage Only at 30-ft
Tile Spacings

DRAINMOD Yield Output for Subirrigation at 30-ft
Tile Spacings

NPV and Gross Margins - Base Weather Sequence (1958-87)
NPV of WTMS Options over the Planning Horizon (1958-87)
Expected NPV, SD of NPV, and Gross Margins Over DR60
Comparison of NPV and ENPV

Yield Sensitivity Analysis

Financial Parameter Sensitivity Analysis

Output Price Sensitivity Analysis

Investment Cost Sensitivity Analysis

83

100

102

104

106

112

113

121

122

124

125

Figure 2.1.
Figure 2.2.

Figure 3.1.

Figure 3.2.

Figure 3.3.
Figure 3.4.
Figure 3.5.
Figure 3.6.
Figure 3.7.
Figure 3.8.

Figure 4.1.

Figure 4.2.

Figure 4.3.

Figure 4.4.

Figure 4.5.
Figure 4.6.

Figure 4.7.

LIST OF FIGURES
Profile and Overhead View of a Drainage-Subirrigation System
Schematic of Water Table Management

Plot of Genesee Historic Yields and Estimated Yield with
Residual Plot

Plot of Reese Farm Historic Yields and Estimated Yield with
Residual Plot

Comparison of Genesee Yields with DM Predicted Yields (*110)
Comparison of Reese Farm Level Yields with DM YR (*130)
Probability Density Function

Cumulative Distribution Function

Cumulative Distributions - FSD and SSD

Cumulative Distributions with the X-Axis and Y-Axis Reversed

DRAINMOD Predicted Yields for Drainage Only and
Subirrigation

Cumulative NPV for Drainage Only WTMS - Base
Weather Sequence (1958-97)

Cumulative NPV for DR60 and Surface Water Subirrigation
WTMS - Base Weather

Cumulative NPV for DR60 and Well Water Subirrigation
WTMS - Base Weather

Histogram of NPV for DR60
Histogram of NPV for SIGOW

Histogram of NPV for SI30S

12

67

68

88

88

91

95

103

108

109

110

115

115

116

Figure 4.8.
Figure 4.9.
Figure 4.10.

Figure 4.11.

Histogram of NPV for $1608
CDFS of Drainage Only WTMSS
CDFS of DR60 and Surface Water Subirrigation Systems

CDFS of DR60 and Well Water Subirrigation Systems

116

118

118

119

CHAPTER 1

INTRODUCTION

1.1. Introduction

Excessive or deficient soil water conditions are potentially the most limiting
factors in corn production. For more than a century, farmers in Michigan and other
states have used surface and subsurface drainage to manage excessive water. Drainage
allows farmers to begin planting operations earlier in the spring and ensures that they
can begin harvest in a timely manner in the fall. In addition, it allows them to remove
excessive soil water to ensure a suitable environment for crop grth during the growing
season.

In the Saginaw Bay Area of Michigan, many farmers are improving their drainage
systems by reducing the Spacing between tiles to benefit from improved drainage. Some
are also adapting their drainage systems to serve a dual role as a subsurface irrigation
system, or subirrigation system. A subirrigation system is a drainage system that has
been modified by installing control structures and irrigation risers and developing a water
supply system to pump water into the drainage tiles during the growing season to supply
supplementary water to crops. Retrofitting a subsurface drainage system for
subirrigation often also entails reducing the tile spacing in order to ensure that the water

table can be maintained at a more uniform level in the field.

2

Through this new water table management system (WTMS) approach, farmers
can not only remove water from fields under excessive water conditions, but can also
pump water back into the drainage tiles and maintain adequate soil moisture throughout
the growing season. Their ability to manage the water availability conditions in their
fields under both excessive and deficient soil water conditions allows them to control
much of the yield risk they face.

Investing in an improved WTMS has risks of its own, however. Making any
investment decision involves financial risk. The farmer must determine if the increased
yield benefit or reduced yield variability from improved drainage or subirrigation
provides enough additional revenue or stabilizes revenue sufficiently to justify the
investment cost. But in humid climates, these benefits are very dependent on the rainfall
pattern in the years following the investment. For a drainage-only system, in most years
the tile spacing might be adequate, but in particularly wet years, improved drainage
might mean the difference between meeting and not meeting planting time constraints.
Similarly, with subirrigation in humid regions, rainfall is adequate in many years to
produce acceptable yields of major field crops grown in Michigan: corn, dry beans, sugar
beets, and soybeans. Without taking the financial risk of investing in irrigation, farmers
can still produce these crops. For a given planning horizon, the profitability of both
types of improved WTMS investment over the existing drainage-onlyl system hinges on
the particular pattern of rainfall following the investment. For example, with
subirrigation if rainfall is adequate in the first few years following the investment, the
additional yield benefit of the system is small and the payback period is lengthened,

which greatly lowers the net present value (NPV) of the investment. If, on the other

 

‘ “Drainage only“ refers to conventional subsurface drainage.

3

hand, rainfall is poor following the investment, the additional yield benefit is large, and
the system pays for itself more quickly, resulting in a larger NPV for the entire planning
horizon.

The question facing farmers is how to assess these issues and make the right
investment choice. Past economic analyses of subirrigation and drainage have provided
some measures of the net returns to subirrigation or improved drainage under actual
field conditions for a limited number of years’ field data and under simulated conditions
for more extensive time periods, providing decision makers with an idea of the economic
benefit of the WTMS investment for a particular sequence of weather. But they have
not provided answers to the larger question of what are the expected returns to an
investment in a WTMS given other possible sequences of weather. The present analysis
attempts to answer that question.

In addition, past economic studies of WTMS have presented results in terms of
mean values but have not adequately assessed whether subirrigation or improved
drainage reduces income variability, and if so, how this benefit should be quantified for
risk averse farmers. This study looks Specifically at the risk implications of investing in a
WTMS for farmers with varying risk attitudes.

Huron County, in the Saginaw Bay area of Michigan, is the hypothetical Site of
this economic analysis. A major impetus for choosing Huron County as the setting for
the analysis is that county farmers are currently installing improved drainage and
subirrigation systems. In addition, studies of subirrigation potential have identified the
five counties of the Saginaw Bay area as having the greatest concentration of acres with
high subirrigation potential in Michigan (Belcher, 1990a). Of those five counties, Huron

County has the largest number of acres of land suitable for subirrigation. On the other

4
hand, it also has limited ground and surface water available for further expanding

subirrigation. Already in some rural townships, ordinances have been passed to limit
groundwater pumping for subirrigation.

This limitation has framed the context of the current analysis. If subirrigation is
to continue to expand in Huron County, water sources other than groundwater need to
be developed. The Saginaw Bay provides a vast supply of potential water for irrigation.
If irrigation districts could be established, the large irrigation potential of Huron County
could be tapped. But subirrigation would not only have to provide adequate on-farm
benefits to offset on-farm costs, it would have to offset irrigation district development,
maintenance, and operating costs. A preliminary study of the economic and technical
feasibility of establishing an irrigation district to draw water from the Saginaw Bay to
bring water to farmers in areas that are particularly suitable for subirrigation has shown
that water costs to farmers in such an irrigation district could range between $25 and $35
per acre (Williams et al., 1990). This cost would be additional to on-farm costs of
developing a subirrigation system and pumping water from the district irrigation canal to
the farmer’s fields. The present economic analysis should provide a measure of the on-
farm benefit of subirrigation over drainage only that could be used as a benchmark for
what farmers might be willing to pay to participate in an irrigation district.

Because some farmers do have access to ground water, the economic analysis is
done for both a well water source and a surface water source. The surface water
investment and operating costs mimic the on-farm costs that a farmer would experience
if pumping from a private surface water source or an irrigation district canal. Operating
and investment costs for a well water source are substantially higher than for a surface

water source because the well drilling cost has to be considered as part of the

5

subirrigation WTMS investment. The results for the well water source subirrigation
investment could provide farmers who are considering developing a private well or
participating in an irrigation district with a measure of the return from each option.

The analysis begins with the assumption that continuous corn is being grown on a
40 acre field that has a Kilmanagh soil and an existing drainage system with tiles spaced
at 60-ft intervals. The strategies evaluated include modifying the existing drainage-only
WTMS by reducing the drain spacing to 20 feet (DR20) or 30 feet (DR30), keeping the
existing drainage-only system intact (DR60), converting the existing drainage-only WTMS
into a subirrigation system at 20-ft tile spacings for a surface water source (81205) or a
well water source (SIZOW), or into a subirrigation system at 30-ft tile spacings for a
surface water source (S1308) or a well water source (SI30W), or into a subirrigation

system at 60-ft tile spacings for a surface water source (81608) or a well water source

(SI60W).

1.2. Objectives
The objectives of this study are the following:
(1) Determine the economic benefit of converting the existing drainage-only WTMS to a
drainage-only system at 20- and 30-ft tile spacings.
(2) Determine the economic benefit of converting the existing drainage-only WTMS to a
subirrigation WTMS at 20-, 30-, and 60-ft tile spacings for a surface water source.
(2a) Determine if the benefit of subirrigation with a surface water source is large
enough at the farm level to offset the water use fees if an irrigation district were

established which could provide water to farmers at a charge of $25-$35 / acre.

6

(3) Determine the economic benefit of converting a drainage-only WTMS to a
subirrigation WTMS at 20-, 30-, and 60-ft tile spacings for a well water source.
(3a) Compare the difference in benefit of subirrigation with a well water source
and a surface water source with potential irrigation district water use fees to
determine if a farmer without access to a private surface water source would be

better off drilling a well or participating in the irrigation district.

Two approaches are taken in the analysis. First, the economic analysis is
performed using a given series of yields derived from a simulation model run with
weather that occurred in Flint, Michigan from 1958 to 1990. This is similar to the
approaches taken to date in evaluating subirrigation. Second, random yield sequences
are drawn from the simulation yield data. Results from the two approaches are
compared. Finally, results from the second approach are analyzed both under certainty

and under risk.

1.3. Methods

The analysis proceeds in seven stages.
(1) The production and investment costs associated with growing corn under different
WTMSS are determined.
(2) The simulation model DRAINMOD is used to generate corn yields and irrigation
application amounts for different WTMSS over a 33-year period of historic weather data.
(3) A net present value (NPV) analysis under the base weather sequence is performed
using the cost data and the output of DRAINMOD. The NPVS of the various WTMS

options are compared.

7

(4) Monte Carlo simulation is used to generate probability distributions of NPVS to
capture the effect of weather variability on NPV.
(5) Expected net present values (ENPV) of the various WTMS options are calculated
and compared across systems. The NPVS from the base weather sequence are compared
with the ENPVS.
(6) The probability distributions of NPV generated in the Monte Carlo simulation are
compared in three stages:
a) using expected value-variance (EV) efficiency criteria;
b) using first and second degree stochastic dominance (FSD and SSD) criteria;
c) using stochastic dominance with respect to a function efficiency criteria to
compare those distributions which are not stochastically dominated by FSD or
SSD.
(7) A sensitivity analysis is performed to compare the outcome of the base analysis with
outcomes of analyses run with a range of product prices, investment costs, financial

parameters, and yield assumptions.

CHAPTER2

LITERATURE REVIEW

2.1. Introduction

The focus of this literature review is subirrigation. Drainage is always an integral
component of any subirrigation system and is thus encompassed. It also receives some
attention individually, but subirrigation is a new technology compared with drainage and
hence is highlighted.

Literature on subirrigation can be divided into three main categories: technical
aspects, water quality impacts, and economic feasibility. In addition to economic studies
that look specifically at subirrigation, there is a large body of literature on economic
aspects of irrigation in general. The purpose of this literature review is to summarize
the most important conclusions of studies of technical and water quality issues related to
subirrigation and to focus on the economic studies of subirrigation and irrigation.

The results of a selection of technical and water quality impact studies are
summarized in section 2.2 below. In section 2.3, a selective overview of the findings of
economic analyses of irrigation in humid climates is presented. A more extensive
presentation of the available economic studies of subirrigation follows in section 2.4,
Section 2.5 highlights the Strengths of available economic analyses and the gaps that need
to be filled. These provide the guidelines of the approach to be taken in the current

economic analysis.

2.2. Technical Aspects of Subirrigation

Michigan has 3 million acres of poorly drained agricultural land (USDA, 1982).
Subsurface drainage tiles make many of these acres of poorly drained, high water table
fields productive for agriculture. The necessity of providing drainage and the frequent
use of subsurface drainage tiles as the drainage system of choice make subirrigation
through those same drainage tiles technically feasible. Below, the technical aspects of
subirrigation are described and the soil and land characteristics that make subirrigation
feasible are noted. Based on these characteristics, the areas in Michigan with high
potential are identified. Finally, available water resources are assessed in those areas

with high subirrigation potential.

2.2.1. Subirrigation Defined and Described

Subirrigation is a method of providing supplementary soil moisture to crops. A
subirrigation system is generally a subsurface tile drainage system that has been modified
so that the drainage tiles serve a dual role of removing excess water and supplying
supplementary water to meet crop needs. Figure 2.1. shows the layout of a subsurface
drainage-subirrigation system. The system components include a main water pipe,
perforated laterals, water control structures, and an irrigation intake riser. The main
water pipe carries water from the water source to the laterals during subirrigation and
collects water from the laterals during drainage and carries the water to either a
drainage canal or some other receiving system. The laterals are perforated pipes, usually
of corrugated plastic. Water seeps into them during drainage and out of them during
subirrigation. The control structure houses the weir which when raised or lowered

controls the water table level in the field. The irrigation intake riser receives the

10
Figure 2.1. Profile and Overhead View of a Drainage-Subirrigation System

Overhead View of Drainage/Subirrigation System

WATER PUHFED IN
FROM A WILL

 

 

 

 

 

 

 

 

 

Larrnxt l

’HAIN

2
l SLOP580.5%

s

Q

5 fi—cournct srnucrunr

e

T

a

, -—courneL srnucruae

 

Profile View of Drainage/Subirrigation System

 

 

 

 

 

 

 

 

CONTROL
__ Scene; srnucrun: CONTROL
2 9 - STRUCTURE
- a, ‘ + ~
{MAIN cum gm: g . _
' . Lunar -' ‘

 

LATERAL

     

Source: Skaggs, 1981, p. 8-5.

11

irrigation water from the water source before it is distributed through the main water
pipe to the laterals.

The principle behind drainage-subirrigation systems is to manage the water table
(Figure 2.2.). The water table is lowered during drainage by allowing water to move
freely from the tile laterals into the main and out of the field. This provides trafficable
conditions during planting and harvest and removes excess water after a heavy rainfall
during the growing season. As can be seen in Figure 2.2., drainage between laterals is
slower than directly over the laterals, creating a dome effect on the water table. During
subirrigation, water is pumped from the main into the perforated subsurface drainage
tiles, raising the water table enough to maintain an adequate water supply just below the
root zone of the crop. Soil capillary action and diffusion draw water upward from the
water table into the unsaturated root zone, replenishing water which plants remove
during evapotranspiration (ET). As in the case of drainage, during subirrigation, water
movement into the soil is not always uniform. Over the tiles the water table bulges
upward and between the laterals it scoops downward.

A subirrigation system can be operated in either of two ways. The most common
procedure is to maintain a constant water level elevation in the tile outlet. Water is
periodically pumped into the tiles to replenish water which moves from the drains into
the soil to supply ET demands and seepage losses. A second procedure is to pump
water into the root zone of the soil profile. After pumping is stopped, the water table
level is allowed to fall to some predetermined level before pumping is initiated again

(Skaggs, 1981).

 

12
Figure 2.2. Schematic of Water Table Management

RAINFALL OR ET

liiiiiliiiiiiiiiiiiiili

DEPRESSION STORAGE S = RUNOFF (R0)
- , - ?

    
 
   

 
 
    

m------'

SUBIRRIGATION

 

WATER TABLE

 

 

 

ii
Ii/l/7i7/l/lf/7f/Tlf/II/I/flf/I/fllfl/IU’7T7/7/777

RESTRICTIVE LAYER DEEP SEEPAC-E (05)

Source: Skaggs, 1981, p. 2-1.

2.2.2. Soil and Land Characteristics

Soil properties are among the most important considerations when assessing the
potential for subirrigation (Kittieson et al., 1990a). The most important soil factor in
subirrigation is the presence or absence of a barrier layer within 72 inches of the soil
surface. This barrier is a natural feature of poorly drained soils. It slows the downward
movement of water into the soil and produces a shallow water table. It is the barrier
layer and the Shallow water table that make water table management both necessary and

feasible. Necessary, because drainage tiles must be used to remove excess water from

13

fields for planting and harvest operations when heavy equipment requires trafficable field
conditions. Feasible, because with the barrier, water pumped into the tiles through a
subirrigation system is prevented from draining immediately through the soil layers and
moving beyond reach of the plant roots.

Other soil factors, such as soil texture, permeability in the top 40 inches and in
the 40- to 60-inch layer, depth to bedrock, and depth to the barrier layer determine the
suitability of soil for subirrigation (Kittleson et al., 1990a).

Field Slope also affects the feasibility of subirrigation. Land with a slope greater
than 2% Should not be considered for subirrigation, while land with a Slope of between 0
and 1% has a high potential for subirrigation. Slope is an important determinant
because for subirrigation to be practicable, the water table needs to be kept at a
relatively uniform level throughout the field. If there is too much Slope to a field, several
zones need to be established within the field and a control Stand installed for each zone
(Figures 2.1. and 2.2.). A control stand should be installed in the main subirrigation line
for every 0.5% change in slope in the field. If the field slope is greater than 2%, adding
the extra control stands necessary to maintain a uniform water table level in the field
increases the cost of the subirrigation system.

Based on the above criteria, researchers at Michigan State University’s Institute
of Water Research and at the Soil Conservation Service estimate that 492,192 acres, or
19.9% of the agricultural land, in the five counties of the Saginaw Bay area have high
potential for subirrigation. Huron county has the largest number of acres of highly
suitable land, with 353,234 acres. In the five county area, another 1,497,433 acres

(59.8%) have medium potential. Thus, 80% of the agricultural land in the study area has

14

either high or medium potential for subirrigation based on soil property criteria alone

(Kittleson et al., 1990a).

2.2.3. Water Availability

Irrigation, both sprinkler and subirrigation, has expanded rapidly in Michigan in
the last decade. In Huron County, the hypothetical Site of the present economic analysis,
in the 11-year period between 1978 and 1988 irrigation acreage increased 500% from 438
acres to 2200 acres (Michigan Department of Agriculture, 1990; LeCureux and Booms,
1990a). In 1987, the total irrigated acreage in Michigan was 15,035 acres, up from 8,460
acres in 1978 (US. Department of Commerce, 1978 and 1987). While the total irrigated
acreage for Michigan is still quite low, there has already been concern about increasing
water demand for irrigation in rural townships in the Saginaw Bay Area. Some
townships have established water use ordinances to limit the continuous operation of
high volume irrigation wells. In Huron County, township administrators are considering
ordinances that would require farmers to obtain permits to pump groundwater for
agricultural purposes (Kittieson et al., 1990b).

Several studies have been made of water availability for irrigation in the Saginaw
Bay area. One study showed that over 83% of the high suitability soils and 55% of the
medium suitability soils are located over geologic formations containing no significant
aquifer. In Huron County, which has the largest number of acres of highly suitable land
for subirrigation, only 19,659 of the 324,000 suitable acres are over an aquifer (Kittleson
et al., 1990a). The authors concluded that if major expansion of subirrigation occurs on

high suitability soils using groundwater, shortages of groundwater and /or decreases in

15

groundwater quality will develop almost immediately in most areas (Kittleson et al.,
1990a).

Considering surface water availability, a recent study by the Department of
Resource Development at Michigan State University (He et al., 1991) estimated that
although 68.1% of the land in the Saginaw Bay area that is within two kilometers of
surface water supplies that have year-round water, a maximum of 44,105 acres (2.2% of
the total agricultural land in the Saginaw Bay area) can be irrigated by stream flow in the
watersheds in the five Bay counties for which stream flow data are available. This
estimate is based on three assumptions. First, it is based on irrigation water demand at
the 75% probability level, which means that irrigation demand in a particular year will be
smaller than the calculated value 75% of the time (He et al., 1991, p.13). Second, it is
based on a 75% exceedence flow, which indicates that Stream flow will exceed or equal
the specified value 75% of the time. Third, it is also based on the assumption that
stream water is drawn down to the 95% exceedence level. This is the flow level set by
the National Pollutant Discharge Elimination System (NPDES) for effluent limits. If
withdrawals did in fact occur up to the 95% exceedence level on a regular basis,
however, they would ”seriously degrade the quality of the stream” (He et al., 1991).

The largest source of water in the Saginaw Bay Area is Lake Huron. Agricultural
producers bordering the lake at Saginaw Bay can and do use bay water to irrigate their
crOps. In fact, most growers who have access to lake-level water irrigate their crops
(Spicer, 1990). Currently, the legal use of lake water for irrigation is limited to lands
that are riparian to Lake Huron or to a tributary stream. Access varies depending on
the lake level, which fluctuates as much as 6.59 feet at the extremes (Spicer, 1990). The

Kittieson et al. (1990a) study mentioned previously showed that there are 37,561 acres of

16
highly suitable land within 2.5 miles of Lake Huron, with 34,074 of these acres in Huron

County. Another 72,816 acres of medium suitability soil fall within this same distance,
and 11,861 of these are in Huron County.

The limited availability of suitable stream and ground water and the elevated
costs of developing groundwater sources have stimulated interest in exploring the
feasibility of establishing irrigation districts to draw water from Saginaw Bay. Several
engineering studies of costs of constructing an irrigation district have been commissioned
by the Saginaw Bay Subirrigation/Drainage (SBSD) project (Spicer, 1990; Williamson &
Associates, 1990; Williams et al., 1990).

One study of a proposed 24,000 acre district in Huron County falling within the
Caseville, Lake, McKinley, and Chandler Townships (Williams et al., 1990) estimated
that for a system designed at a capacity to deliver 8 inches of water in 40 days to 50% of
the area farmland, the total annual cost would be between $25 and $35 per acre. These
estimates include amortization of the construction costs and annual operation and
maintenance costs, assuming 8% financing over 20 years. These figures are only for
delivery of the water to a farmer’s field. Once delivered, water must be distributed and
the costs of the on-farm distribution system are additional costs.

A second study (Spicer, 1990) reports that a 2,400 acre Mud Creek Irrigation
District in Huron County is being established to withdraw water from Saginaw Bay.
Their estimates are that the average cost per acre for a district encompassing 5 miles of
land fi'om the bay inland would be $841, which translates into an average annual cost of
$32.30 per acre, based on a 20 year depreciation period and 8% interest. These costs

include construction costs, interest, electricity, and maintenance costs. In this case,

l7

farmers would again incur additional costs to bring the water from the irrigation ditch to
their fields.

The per-acre cost estimates of establishing the irrigation districts necessary to
pump from the bay are relatively high given that farmers must then incur additional costs
to bring the water to their fields. Part of the impetus of the current economic analysis of
subirrigation and others done in Huron County is to determine if the benefits of
subirrigation are large enough at the farm level to cover additional costs of water
brought to farmers through an irrigation district at a per-acre cost of as much as $35.
The results of on-farm economic studies will establish whether irrigation expansion
should be limited to acreage that can be irrigated with surface water (other than lake
water) and ground water, or whether it will be economically viable to establish irrigation

districts and greatly expand irrigated acreage through the use of lake water.

2.2.4. Water Quality Aspects of Subirrigation

Water quality research has focused on whether subirrigation results in more or
less contaminants being discharged into tile effluent, being lost to surface water runoff,
or remaining in the field than conventional subsurface drainage or surface-only drainage
systems. Results are rather mixed. Two researchers at Michigan State University’s
Department of Agricultural Engineering summarized the results of 43 research reports
published in scientific journals and 18 additional research articles in a literature review
of water table management impacts on water quality (Fogiel and Belcher, 1991). They
concluded from the studies that the primary impact of water table management is on
receiving surface waters (as opposed to groundwater) and that in general, subsurface

drainage systems reduce runoff and therefore result in less sediment and fewer pollutants

18

attached to soil particles being delivered to surface waters. However, some studies
showed that improper management of subsurface drainage systems can result in
increased nitrate nitrogen concentrations and loadings being delivered to receiving
waters.

Only two of the studies Fogiel and Belcher reviewed looked Specifically at the
effects of subsurface irrigation (SI) on water quality. One study (Campbell et al., 1985)
compared differences in nitrate-nitrogen and orthophosphate losses on a sandy soil
between a water furrow-irrigation system and a subsurface drainage-irrigation (SI)
system. The authors found that SI system reduced overland flow and sediment loss and
also resulted in reduced nitrogen loading and potassium concentrations (Fogiel and
Belcher, 1991). A Michigan water quality pilot study (Protasiewicz et al., 1988)
comparing nutrient and pesticide loads carried to the edge of field in subsurface drain
flow between a conventional subsurface drainage system and a subirrigation-drainage
system over an 8-month period reported that levels of nitrate-nitrogen and phosphorous
carried to the edge of the field by subsurface drain flow were higher for the conventional
subsurface drainage system than for the subsurface irrigation-drainage system while
levels of potassium and atrazine carried to the edge of field were lower for the
conventional subsurface drainage system.

Field trials in Huron County (LeCureux and Booms, 1990c,d; LeCureux, 1991a,b)
over the 4-year period 1987-90 showed that subirrigation does not contribute additional
amounts of nitrates or pesticides to tile effluent. Some of the data indicated that
subirrigation allowed crops to better utilize nutrients, thus reducing the residual amounts
of these chemicals being lost in the tile system or in surface runoff. Data from 1988, a

low rainfall year, showed that the subirrigated fields released lower levels of nitrates into

19
tile effluent in the fall than the drainage only fields. On the other hand, data from 1989,

a more plentiful rainfall season, indicated that both systems released approximately the
same level of nitrates into the tile effluent.

Fogiel (1992) monitored the effects of water table management on nutrient and
chemical loadings in surface and subsurface runoff and in the soil for 2 years at a field
site in Unionville, Michigan. He found that in comparison with a no subsurface drainage
(NSD) system, both subirrigation (SI) and conventional drainage (DR) reduced surface
drainage outflow and surface drainage orthophosphorus loading for an above average
rainfall (AAR) growing season drainage but resulted in similar surface drainage and
surface drainage loadings of orthophosphorus during a below average rainfall (BAR)
growing season. For both AAR and BAR growing seasons, both SI and DR resulted in
increased total flow from the field but in reduced surface drainage loadings of nitrate
nitrogen and potassium compared with the NSD treatment. In comparing SI and DR,
Fogiel found little difference in effect on surface drainage outflow for either growing
season, but that 81 increased tile outflow and reduced tile outflow loadings of nitrate
nitrogen and orthophosphate phosphorus for the AAR growing season. 81 increased
potassium tile outflow loading for both seasons. Testing of field samples for the top 0.3
meters of soil for alachlor and nutrients Showed no traces of alachlor for any WTMS, but
both tile drained treatments were found to have significantly higher orthophosphate
phosphorus loadings for the AAR season and the SI treatment had Significantly higher
orthophosphate phosphorus than did either the DR and NSD treatments for the BAR
season. For both growing seasons, the tile drained treatments had significantly higher

potassium loadings than the NSD treatment.

20

To summarize these results, in comparing subirrigation with drainage only,
subirrigation results in lower nitrate nitrogen losses in surface runoff and in the tile
effluent and in higher potassium and atrazine loadings in the tile effluent, regardless of
seasonal rainfall. Subirrigation results in higher phosphorus soil concentrations in low
rainfall years.

Because current scientific understanding of the effects of drainage and
subirrigation on water quality is limited and inconclusive, a research site has been
established in the Saginaw Bay area to study these effects. Both nutrient and pesticide
sampling of tile effluent, and surface runoff, and soil will provide additional data to judge

the environmental effects of water table management.

2.3. Economic Analyses of Irrigation

There is a large body of literature on economic analyses of irrigation in general.
These can be divided into those that deal with irrigation application strategies under a
given irrigation system and those that deal with the investment decision concerning which
irrigation system to install given several choices. For the present analysis, the irrigation
investment literature is most relevant.

Because irrigation investment decisions involve a long planning horizon, the
economic analyses in general use some form of net present value (NPV) model as the
analytical tool for evaluating the profitability of the investment. The objective function
ascribed to irrigation managers differs depending on the decision environment. Boggess
et aL (1983) performed a review of all the irrigation Strategy analyses, including
investment strategies, reported in professional journals to determine what Specific

objectives were ascribed to the irrigation manager and how the issue of variability in the

21

decision environment shaped the specification of the objective function. They
summarized the key sources of variability addressed in the studies as follows:

(1) Variability in aboveground conditions, which include plant capabilities, soil

cultivation practices, level of weed control, wind conditions, degree of solar

radiation, rainfall quantity and timing, humidity and temperature.

(2) Variability in below ground conditions, including rooting depth and density,

nutrient movements and levels, water holding and hydraulic features of the soil,

proximity to ground water, and infiltration rates.

(3) Variability of product price.

(4) Variability in marginal costs of irrigation water, including fuel and labor costs

and cost differences related to the design of the irrigation system.

(5) Variability of institutional features of the water supply system, including rules

affecting when water can be pumped, how much can be diverted, and when it can

be used at all.

They concluded that yield variability as influenced by above and below ground
conditions has received the brunt of attention in the literature. Yet they were surprised
by the fact that only three studies (Yaron and Strateener, 1973; Harris and Mapp, 1980;
and Boggess et al., 1981), of 52 studies reviewed, presented estimates of the variance of
profits stemming from yield variability associated with various strategies and only two of
those three (Yaron and Strateener, 1973; Boggess et al., 1981) posited that profit
maximization subject to minimum variance of profits represents a credible goal of
irrigation managers. The other studies used single-dimensional decision criteria to

determine optimal irrigation strategies. The objective functions included yield

22

maximization, profit maximization, water use minimization, yield maximization given a
fixed quantity of water, etc.

To illustrate the diversity of approaches emphasized by Boggess et al. (1983), two
studies are presented below. The Wilson and Eidman study (1983) is a straight-forward
investment analysis that assumes the goal of the irrigation manager is to choose the
irrigation strategy that maximizes profits under conditions of certainty. It assumes
known and fixed financial parameters and uses an average yield differential between
irrigated and nonirrigated production. The Boggess and Amerling study (1983) begins
with the assumption that the goal of an irrigation manager is to choose a production
strategt that reduces income variability. It therefore tries to assess the impact of
variations in weather patterns and the associated variability in yield differentials between
irrigated and dryland production on the profitability of irrigation investments in humid
regions.

Wilson and Eidman ( 1983) used the results of a survey of irrigators in the
southwest and south central regions of Minnesota to perform a financial analysis of
investing in a center pivot irrigation system for several different soils found in the two
regions. The impetus for the study was that farmers in the aftermath of the 1974-76
drought installed irrigation systems without having specific information about the
financial profitability of irrigation for their soils.

The authors obtained information from irrigators for both irrigated and
nonirrigated yields of corn and soybean and developed a yield differential model for
predicting corn yields. They used this information to analyze the profitability of
investing in a center pivot irrigation system for different soils using an after-tax net

present value model. In their analysis, they abstracted from all potential sources of risk,

23

assuming an average yield differential over the 15-year planning horizon, assuming fixed
investment costs, fixed crop prices, and fixed pumping costs each year. They found that
irrigation is profitable on soils with a moderate available water capacity (lighter soils) but
is not profitable (assuming a 12% desired after-tax rate of return) on high and very high
available water capacity soils (heavy soils). The lower profitability on the heavier soils
was directly related to the lower yield differential between dryland and irrigated yields
on these soils.

Boggess and Amerling (1983) investigated the importance of the pattern of
weather variability on irrigation investment decisions in humid climates. They argued
that irrigation investment decisions in humid climates differ markedly from those in arid
climates because of the sensitivity of net present value (NPV) to the particular weather
sequence over the economic life of the investment. They were able to capture this
particular feature of the investment decision by using Monte Carlo techniques to
generate probability distributions of NPVS. They used a simulation model to generate
dry-land and irrigated crop yields based on a 17-year time series of historical weather
data and incorporated the results from the simulation model into a net present value
analysis. For two different soil groups and three different crops (corn, peanuts, and
soybeans), they studied the profitability of four alternative irrigation investment options,
two based on a center pivot system and two based on a traveling gun irrigation system.
They found that the profitability of the systems over a dryland production system
difiered by crop and by soil type. For a sandy soil, only the low pressure center pivot

(LPCP) irrigation system had a positive expected NPV for all three crops1 and peanuts

 

‘ NPV figures were reported as the additional benefits of investing in irrigation as compared to
dryland production of the same crop.

24

was the only crop for which all four irrigation systems had positive expected NPVS. For
a sandy loam soil, none of the irrigation systems had a positive expected NPV.

Sensitivity analysis of the NPV results to marginal tax rate, inflation rate, product
price, and yield response for the LPCP system revealed that the expected NPV of
investing in irrigation is relatively more sensitive to yield response and output prices than
to the other parameters. The authors found that of the three crops studied, the NPV
results for corn, which had the lowest per-unit value of the three crops, were the most
sensitive to yield response. This finding highlights the fact that the assumed level of
product price can have a significant impact on the NPV results and that sensitivity
analyses should be done for a range of product prices.

For the sandy soils, the authors compared the cumulative distribution functions
(CDFS) of NPV for the different investment options to determine stochastic dominance.
They found that within crops, the systems could be ordered by first degree stochastic
dominance (FSD), with the LPCP system dominating the others. This particular result
occurred because all the irrigation systems were assumed to produce equal yield results
but involved differing investment costs. In comparing the CDFS of NPVS across crops on
both soil types for the dominant irrigation system, the LPCP system, they were able to
set priorities for irrigating different crops given either soil type. Thus they found that on
sands, peanuts dominate corn by FSD and both dominate soybeans by FSD. On sandy
loams, they found that com dominates both soybeans and peanuts by FSD, while

soybeans dominate peanuts by SSD.

25

2.4. Economic Evaluations and Studies of Yield with Subirrigation

Unfortunately, literature on the economics of subirrigation is limited. A
thorough literature search revealed no studies done by economists on subirrigation.
Available economic and yield studies fall into two categories: Those done by extension
agents and engineers using actual data from field trials and those done by engineers
using simulated data. The studies will be summarized below based on this distinction.

Before going into the results of specific economic analyses of subirrigation, some
general information is provided first to clarify basic cost considerations in water table
management systems. For any drainage (DR) or subirrigation (SI) system, cost
effectiveness depends on the crop, soil, topography, climate, water supply (for SI), and
degree of management. The two main expenses involved in installing and operating a
water table management system are the costs of providing a water supply and of
installing underground tiles. Tiles must be laid for both DR and SI systems, but for
effective subirrigation, the tile spacing is often narrower. The tiles and water supply
costs are very site specific and can vary as much as 300%. Evans et al. (1988) give a
very detailed analysis of costs associated with different drain spacings, water sources,
control systems, and pumps for North Carolina. The cost of a water supply varies
greatly by water source. For a surface water source such as a pond, stream, lake, etc.,
the initial investment cost is limited to the pump and electricity hookup charges. If a
surface water source is not available, the SI investment costs include the well installation
costs and higher pump investment costs. In addition, Operating costs for a well water
source are higher than for a surface water source. In the Huron County studies
presented below (LeCureux and Booms, 1990a-d; LeCureux 1991a,b) electricity costs for

surface pumping ranged from $4.19 to 3560/ acre, while those for deep well pumping

26
ranged from $9.60 to $12.35/acre, a difference of $5.41 to $6.75/acre. Pump

depreciation and interest costs for surface pumping were reported as $15.35 / acre based
on a 7-year depreciation period and 12% interest. The same costs for deep well
pumping were $32.57/acre. The combined difference between these system costs for a
well water source and a surface water source is as high as $24/acre.

The reality of these cost differences is evident in the water source use patterns of
agricultural producers in the Saginaw Bay area. A 1988 subirrigation inventory survey by
Belcher and Wood (1990) found that 90% of survey respondents used a surface water
source such as a stream or river (30%), a ditch (43%), or a pond (17%). Only 13% of
the respondents reported using a well as their water source.2 That same inventory
revealed that 69% of the subirrigation systems were originally drainage only systems that
were retrofitted for subirrigation.

From the above, it is clear that any economic analysis must clearly specify the
underlying assumptions about costs and the characteristics of the site. More general
conclusions can be drawn from economic analyses if sensitivity analyses of key
assumptions are performed to give a better idea of the range of outcomes possible when

a key parameter is changed.

2.4.1. Field Studies
2.4.1.1. Yield Analyses
Michigan-specific yield data for subirrigation is available through two unpublished

theses (Fogiel, 1992; Belcher, 1990). Fogiel’s study evaluated the effect of subirrigation

 

’ Some respondents used both a well and a surface water source so the total does not equal
100%.

27
on corn yield at a field site in Unionville in Tuscola County during 1990 and 1991. All

management inputs except water table management were the same for three treatments:
a subirrigated treatment (SI), a drainage-only treatment (DR), and a no-drainage
treatment (ND). The results Show that SI yields were 9% higher than DR yields and
17% higher than ND yields in 1990, a year where rainfall was 32% above the 30-year
average rainfall for the site. In 1991, a year where rainfall was 52% below the 30-year
average, the SI yield was 58% higher than the DR yield and 76% higher the ND yield.
Fogiel’s results show that subirrigation produces a significant yield benefit over no
drainage in both wet and dry years, with the largest increase derived in dry years, and
produces a modest yield benefit over drainage only in wet years, but a substantial yield
increase in dry years.

Belcher (1990) studied the yield effect on corn and soybean of varying the water
table level at two field sites: St. Johns in Clinton County and Bannister in Gratiot
County. At the Bannister site, which has a Ziegenfuss soil, subirrigation tile laterals
were spaced 20-, 40-, and 60-ft apart and at the St. Johns site, which has a Wasepi soil,
they were spaced 40-, 56-, and 79-ft apart. Corn yield results at the Bannister site for
1986, the first year after the system was installed, at the different Spacings and water
table depths ranging from 38 to 95 cm, did not Show a high correlation between water
table depth and yield nor were there noticeable yield differences between the different
tile spacings. These results could have been due to the fact that the soil at the site had
been disturbed the previous season when the subirrigation system was installed.

In 1987, water table depths were allowed to vary more significantly than in 1986
(from 48 to 158 cm) and there was a more noticeable treatment effect between the

different water table depths at the same tile spacing. At the 20-ft tile spacing, the corn

28

yield was 226 bushels/acre (bu / acre) at a water table depth of 48 cm compared with a
yield of 138 bu/ acre at 63 cm and 172 bu/acre at a 144 cm water table depth. At the 60-
ft tile spacing, yields increased steadily with decreases in the water table depth. At a
depth of 123 cm the yield was 138 bu/acre, at 96 cm it was 156 bu/acre, and at 82 cm it
was 200 bu/ acre.

The St. Johns results are Similarly mixed for 1987, the year following installation
of the system. However, in 1988, a very dry year, the relative yields increased at each
tile spacing as the water table depth was raised. At the 40-ft tile spacing, corn yields
rose from 116 bu/acre to 166 bu/ acre as the water table depth was raised from 112 cm
to 70 cm. Similarly, at the 80-ft tile spacing, relative yields rose from 116 bu/ acre to 182
bu/acre as the water table depth was raised from 112 cm to 71 cm.

Other field studies of yield for subirrigated crops outside of Michigan are
available and provide useful information for comparative purposes. Most of these
studies use the results of field data from other researchers to validate the DRAINMOD
simulation model (Hardjoamidjojo et al., 1982; Hardjoamidjojo and Skaggs, 1982).

Hardjoamidjojo et al. (1982) studied the effect of drainage, including surface
drainage, tile drainage, and a combination of tile and surface drainage, on corn yields
under excessive soil water conditions. Their findings are important because they
highlight the fact that improving drainage alone on poorly drained fields greatly improves
corn yields.

In reporting their results, Hardjoamidjojo et al. used the concept of relative
yields. This concept warrants explanation here because it will appear again when results
of simulation studies are summarized. Relative yield in percent terms (YR) is the actual

measured yield (Y) divided by the potential yield (Yo): YR = Y/Yo x 100.

29
The potential yield can be either the highest yield obtained in a particular field

trial or it can be the yield goal set based on current technology and given perfect growing
conditions. The concept of relative yield is used to eliminate the effects of factors other
than the drainage treatment effects when comparing yields across trials and across years.
Hardjoamidjojo et al. (1982) ran field trials in Ohio under excessive soil water
conditions from 1962-64, 1967-71, 1976-79 and compared the effect of no drainage,
surface drainage only, tile drainage only, and tile plus surface drainage treatments on
corn yields. Excess water was the only stress the corn plants were exposed to. Under
these conditions relative yields improved consistently from a Situation of no drainage, to
surface drainage, to tile drainage, to tile plus surface drainage. With tile plus surface

drainage, relative yields were greater than 90% in 8 out of the 13 years.

2.4.1.2. Economic Analyses

In Michigan several economic studies of subirrigation on row crops have been
done by an extension agent, Jim LeCureux, working with farmer cooperators in the
Saginaw Bay area as part of the SBSD project. These Huron County studies are
published in two volumes (D’Itri and Kubitz, 1990 and 1991) and report the results of
field trials from 1987-90. The crops evaluated include corn, soybean, sugar beets, and
dry beans. LeCureux evaluated subirrigation for crops singly and as part of a rotation
scheme (LeCureux and Booms, 1990a-d; «LeCureux, 1991a,b). In addition, two studies of
the economies of subirrigation on alfalfa have also been done in the Saginaw Bay area
(Auernhamer and Belcher, 1990; Protasiewicz and Auernhamer, 1991).

Because of the relevance of the Huron County studies to the present study, a

fairly comprehensive accounting is made of the approach taken in both the trials and the

30

economic analysis of the trials. The results of the 1987-90 field trials are summarized in
Tables 2.1 - 2.5. In order to present as much information as possible in the tables,

abbreviations have been used extensively. These are summarized below.

 

AthRViaticn Simifisance

Mgt Management Strategy

Drain Space Spacing between drainage tiles

SI Adv Difference between results of the subirrigated treatment and the

companion drainage only treatment for the same tile spacing
CN Corn

SB Soybeans

NB Navy beans

SBT Sugar beets

WH Wheat

SI Subirrigated

HW High Water Table Goal (close to soil surface)
MW Medium Water Table Goal

LW Low Water Table Goal (further from soil surface)
DR Drainage only

R30 30-inch crop row spacing

R15 lS-inch crop row spacing

1987 Field Trial Results (See Table 2.1)

The 1987 field trials (LeCureux and Booms, 1990a) evaluated the effect of
subirrigation on corn yields at two Sites where the tile spacings were 25 feet on a
Kilmanagh soil (Site 1) and 60 feet on a Shebeon/Kilmanagh soil (Site 2). At both sites
subirrigated treatments (81) were compared to drainage only treatments (DR).

At Site 1, two water table levels were tested for the subirrigated plot and
alternate yield goals of 200 bu/ac, 180 bu/ ac, and 160 bu /ac were set for the irrigated
zones. A single yield goal of 160 bu/ac was set for the DR treatment. Plant populations
and fertilizer levels varied by yield goal. The logic behind establishing three different

yield goals for the subirrigated zone and a single lower yield goal for the drainage only

31
zone is that higher yields can be expected for the subirrigated plots because water is not

a limiting factor. Thus, separate production functions are assumed for the different
treatments. At Site 2, a single yield goal of 180 bu/acre was established for both

treabnenut

Yield results for the different zones at both sites are summarized in Table 2.1.

TABLE 2.1: Summary of Yield Results for 19873

 

 

 

 

 

Crop Mgt Drain Yield Actual 81
Space Goal Yield Adv
CN SI 25
HW 200 174 59
180 170 55
160 155 40
LW 200 176 61
180 171 56
160 160 45
DR 25 160 115
CR SI 60 180 121 0
DR 60 180 121

 

 

 

 

 

 

 

lieéurefix an: Booms, 1393a)

(Source:

Because the 1987 economic analyses do not include depreciation costs or interest
for the pumping equipment and other control structures necessary for subirrigation, net
revenue figures are not included in Table 2.1. These costs were included in all
subsequent Huron County economic analyses and results are therefore presented and
compared for the later analyses.

The results at Site 1, where a well water source was used, Show that in a year
where the rainfall is unevenly distributed, such as 1987, subirrigation produced a yield

difference of as much as 62 bu/ acre over drainage only.

 

3 Seasonal rainfall for 1987 at Site 1 was 16.25 inches and at Site 2 it was 14.65 inches. Normal
seasonal average rainfall for Huron County is 17.3 inches.

32

At Site 2, the surface water source used for subirrigation could not provide
adequate water to subirrigate the test plot. In addition, at the 60-ft tile spacing it was
not possible to maintain a uniform water table in the subirrigated field. Under these
conditions, there was no yield difference between the partially subirrigated plot and the

drainage only plot.

1988 Field Trial Results (See Table 2.2)

The second year of field trials, 1988, was also a dry year in Huron County. Corn
trials were repeated at Site 1 and Site 2 (although on different plots) and a sugar beet
trial was run at a third Site, Site 3, which also had a Kilmanagh soil (LeCureux and
Booms, 1990b-d). At Site 2, a different field from the 1987 one was chosen for the
field trial. In this field, the tiles were spaced at 25 feet instead of 60 feet. Instead of a
subirrigated and drainage only comparison, this time three water table depths were
compared: 12-inch (HW), 18-inch (MWl and MW2), and 35-inch (LWl and LW2) and
three yield goals (160, 180, and 200 bu /ac) were established. One area of the low water
table field had water levels that remained at 48 inches and yields from this plot are
considered equivalent to a drainage only treatment for comparisons.

The gross margin analysis for subirrigation included annual principal and interest
payments on the pump installation and material based on 12% interest over 7 years and
annual per acre electrical charges for pumping water. Seed and fertilizer costs for
subirrigated corn were higher than for the drainage only case using the same assumption
as before that a lower yield goal needs to be set for the drainage only treatment.

The results of the trials and the net margin analysis are summarized in Table 2.2.

For the economic analysis, LeCureux compared the HW SI treatment with the DR

33

treatment and found a yield advantage of 47 bu/acre and a net revenue advantage of
$90/acre for the subirrigated corn. Although LeCureux did not do an economic analysis
of the different water table level management schemes, the yield results Show that there
was a difierence in yields due to the variation in water table levels. Yields for the field
where the water table averaged 12 to 18 inches were 30 bushels higher than yields for
the 35-inch water table field. However, there was very little difference in yields between
the 12- and 18-inch water table treatments.

At Site 1 alternate water table levels were established in the subirrigated zone
and a drainage only zone served as a control. All treatments had 25-ft tile spacings. A
yield goal of 160 bu/ acre was set for the drainage only treatment and less fertilizer and
seed were used. A 180 bu/acre yield goal was set for the subirrigated plots.

Yield results Show no significant yield difference between high and low water
table zones, but a significant 71 bu/acre difference between the subirrigated treatments
and the drainage only treatment. The economic analysis Shows a $72/ acre benefit to
subirrigation (Table 2.2). The smaller gross margin at Site 1 compared with Site 2,
where the yield difference between SI and DR corn was smaller than at Site 1 was due
to the higher cost of well water compared with surface water. In both cases, however,
the results Show that no extra benefit is derived from maintaining too high a water table.

A sugar beet trial was run at Site 3. Two zones with 60-ft and 30-ft tile spacings
were divided into subirrigated and drainage only plots. As with the previous trials, the
subirrigated treatment received different levels of inputs (high /low nitrogen and
high/low plant populations) to reflect differing yield goals, while the drainage only
treatment received only the lower level of inputs (low nitrogen / low plant population). A

summarized version of the results is presented in Table 2.2.

34

In his economic analysis, LeCureux compares the yields and net revenues of the
SI and DR treatments for the 30-ft tile spacings and reports a $158.48/acre difference in
net revenue in favor of the subirrigated treatment.

In the published results, LeCureux notes that there is a difference in yields
between the 30-ft and 60-ft tile spacings for both the subirrigated and drainage only plots
(Table 2.2). He does not, however, evaluate the economic benefit of splitting the tiles.
In Michigan, many farmers are already ”splitting tiles,” i.e., decreasing the distance

between drainage tiles, to benefit from the increased yields of improved drainage.

TABLE 2.2: Summary of Field Trials and Economic Analyses for 1988‘

 

 

 

 

 

 

 

 

 

 

 

Actual SI Net 81
Yield Adv Rev5 Adv
151 71 $174 $72
151 71
90 $102
155 47 $273 $90
154 45
154 45
133 14
132 13
119 $183
22.1 5.5
24.5 5.4 $487 $158
15.5

2 18.1 $329

 

 

 

Sor_~ce._“ . _ Booms, ° ‘ c,d)

 

‘ Seasonal rainfall in 1988 at Site 1 was 16.25 inches, at Site 2 it was 10.02 inches, and at Site 3
it was 11.3 inches.

’ Corn price I: $2.50/bu. Sugar beet price = $31.00/ton.

° Refers to the crop planted previously at the Site.

 

35

Using the yield and cost data provided by LeCureux, a rough gross margin
analysis of the benefit of splitting the tiles for this trial can be made for both the
subirrigated and drainage only cases. The retrofit calculations were made using an
annual principal and interest payments figure of $39.12 on tile and material. This figure
is based on 12% interest over 15 years (LeCureux and Booms, 1990c, p.235). Using this
figure, the benefit to Splitting the tile for the subirrigated treatments is $52.68 and for

the drainage only treatment is $33.11!: $17.92 (Table 2.3A, 2.3B).

TABLE 2.3A: Benefit to Splitting Tiles for Sugar Beet - SI Treatment

 

 
 

ECONOMIC INFORMATION: SITE 3

 

  

 

  

 

  

 

 

 

  

 

  

 

  

 

  

 

   

30-ft tile space 60-ft tile space
Treatment SI SI
Yield 24.5 tons 22.1 tons
Gross Income ($31/ton) $759.50/acre $658.10/acre
EXPENSES (s):
. ($4/ton) 5 98.00 5 88.40
' Tile 5 39.12
TOTAL EXPENSES $137.12 $ 88.40
N GROSS MARGIN $622.38 $569.70

 

 

 

   
   
 

DIFFERENCE: $ 52.68

LeCrex a7 1 c,d)

    
 

 

( see "

 

36

TABLE 2.3B: Benefit to Splitting Tiles for Sugar Beet - DR Treatment

 

 

ECONOMIC INFORMATION: SITE 3

 

30-ft tile space

60-ft tile space

 

 

 

Treatment DR DR
I Yield 18.1 tons 15.5 tone
I Gross Income (SBl/ton) $ 516.10/acre S 480.50/acre

 

 

  

expenses (5):

 

 

 

 

 

 

($4jton) $ 72.40 S 62.00

$ 39.12
TOTAL EXPENSES $ 115.52 5 62.00
GROSS HARGIN S 400.58 $ 418.50

 

 

I DIFFERENCE: (5 17.92)
(Source: LeCureux an Booms, 1 90c, )

 

This simple example shows that under the climatic, management, and cost

conditions prevailing in 1988, splitting tiles for drainage only when growing sugar beets

was not economically attractive. At the same time is shows that splitting the tiles for

subirrigation under 1988 conditions was profitable.

1989 Field Trial Results

Beginning in 1989, LeCureux extended his analysis to include the effects of a

rotation (LeCureux, 1991a,b). Agricultural producers in the area rotate some

combination of corn, dry beans, soybeans, and sugar beets. LeCureux tried to capture

the significance of this practice for the economic viability of subirrigation.

In order to utilize all the available yield data from the various trials, LeCureux

averaged the yield and net revenue results of all the 1987 and 1988 corn and sugar beet

trials with the results of the 1989 corn, soybean, navy bean, and sugar beet trials (to be

discussed below) and then compared the subirrigation and drainage only yield and net

37

revenue figures. For the corn, beet, soybean, navy bean rotation, he reports an average
benefit to subirrigation of $24.37/ acre.

This approach utilizes all the available data, but it does not give an accurate
picture of the benefit to subirrigation under rotation at the level of a single field. From
the 1988 and 1989 field trials, data were available for a corn - soybean rotation at Site 2
for the same field and on a sugar beet - corn rotation at Site 3 for the same field. These
data can be used to consider the economic benefit to subirrigation under a rotation on
the same field. This approach will be illustrated below after the results of the 1989 field
trials are summarized.

One other issue that LeCureux addressed in analyzing the 1987-89 results of the
various trials was whether subirrigation reduces year-to-year yield variability for the
different crops. He again compared results from all the field trials at the different farms.
Again, ideally, an analysis of variability should be done for the same field with the same
tile spacing. LeCureux did not have a long enough data series to make such an analysis.

In 1989, LeCureux ran three field trials at Site 2: sugar beets, soybeans, and navy
beans. The sugar beet trial was run in a field with 30-ft tile spacings that was
subirrigated and in a field with 60-ft tile spacings that had both a subirrigated and
drainage only treatment. All plots received the same level of inputs.

The results of the yield and economic analysis are presented in Table 2.4.
Comparing results for the 60-ft tile spacing, the subirrigated treatment had a $33.05 net
benefit over the drainage only treatment. Comparing the 30-ft SI treatment and the 60-
ft SI treatment to evaluate the benefit of splitting the tile for subirrigation, we see that
the benefit is only $2/acre. Using cost data from the 1988 analysis, where costs for

splitting tiles were reported as $39.12/acre based on 12% interest over 15 years, it is

38

clear that under 1989 conditions it the economic benefit of splitting the tiles for
subirrigation is negative.

Because a 30-ft drainage only treatment was not included in the trial, we cannot
evaluate the benefit to splitting tiles for drainage only. This would be a useful
comparison, however, because some of the yield benefit of the 30-ft SI treatment over
the 60-ft DR treatment might actually be attributable to improved drainage alone in
moving to a smaller drain spacing. For sugar beets, the yield benefit of the 30-ft SI
(subirrigated) treatment over the 60-ft DR (drainage only) treatment was less than 2
tons. Comparing the economic results of the 30-ft SI treatment and the 60-ft DR
treatment, a benefit of $34.88 is gained from bo_th splitting tiles and subirrigating. As
mentioned above, if the costs of splitting tiles are taken into consideration, the benefit of
splitting tiles and retrofitting a drainage only system for subirrigation is negative. These
conclusions, of course, depend on the price of beets, the level of management, and the
rainfall for the 1989 season. Rainfall in 1989, while still below average, was more
plentiful and much more evenly distributed than in 1987 or 1988. The 1988 beet trial at
Site 2 Showed a greater than 6 ton yield increase due to subirrigation, whereas in 1989
the yield increase was less than 2 tons. This example illustrates how sensitive the
economic benefit to subirrigation is to seasonal rainfall, crop price, and crop gown.

The 1989 corn trial at Site 3 was run on the same field as the 1988 sugar beet
trial. As before, both the 60-ft tile spacing and 30-ft tile spacing fields were divided into
subirrigated and drainage only plots. The management scheme in this trial included
using different corn varieties to evaluate sensitivity of results to plant variety and
applying different nitrogen levels to test sensitivity of yields to nitrogen level. The

management, yield, and net revenue results are summarized in Table 2.4. Only the low

39

nitrogen regime was applied to the drainage only treatment. Yield differences by plant
variety and level of nitrogen were insignificant.

While the yield results show a positive benefit to subirrigation in all cases, they
also show a negative benefit in terms of net revenue. Any increase in yield was offset by
the additional costs of production (pumping charges and equipment depreciation,
additional fertilizer and seed costs) associated with the subirrigated treatments. In this
trial, the largest net revenue was for the 30-ft tile spacing DR treatment. As in the beet
trial for 1989, these results demonstrate that the economic benefit to subirrigation is
largely dependent on seasonal rainfall. In good rainfall years, subirrigation produces
little or no economic benefit, whereas in poor rainfall years, the benefit can be

substantial.

EN...

EM

PfiQ

40

TABLE 2.4: Summary of Field Trials and Economic Analyses for 19897

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SI
Adv
533 i
(ON) R30 40.2 0.4 $147 ($17)
R15 40.4 0.3 $144 (518)
DR
R30 39.8 $164
R15 40.1 $162
SB #2 81
(CN) R30 38.6 0.9 $135 ($15)
R15 38.4 (0.5) $129 ($23)
DR
R30 37.7 $150
R15 38.9 $152
2 NB #1 SI 25
(ON) R30 20.0 (0.9) $390 ($37)
R15 21.6 (0.6) $416 ($34)
DR
R30 20.9 $427
R15 22.2 $450
=## fl
3 CN #1 SI 60 200 180.6 27.1 $172 $16
(SBT) 175.1 21.6 $151 ($ 5)
170.1 16.6 $154 (5 2)
30 200 183.6 15.2 $174 ($11)
182.3 13.9 $168 ($17)
178.3 9.9 $167 ($18)
DR 60 200 153.5 $156
30 200 168.4 $185
CN #2 SI 60 200 158.5 3.3 $127 ($36)
(SBT) 173.6 18.4 $156 (5 7)
164.1 8.9 $153 ($10)
'30 200 178.7 14.3 $161 ($17)
180.6 16.2 $168 ($10)
171.4 7.0 $168 ($10)
DR 60 200 155.2 5163
30 200 164.4 $178

 

_
(Source: LeCureux, 1991a)

 

7 Seasonal rainfall in 1989 at Site 2 was 12.46 inches and at Site 3 it was 10.42 inches.

' Corn Price = $2.30/bu. Sugar Beet Price = $40.00/ton. Soybean Price = $5.75/bu. Navy Bean
Price = $24.00/mt.

9 Refers to a specific variety when more than one variety was used in the trial.

41
The 1989 soybean and dry bean trials at Site 2 can be briefly summarized as

follows: Soybeans were gown on a 25-ft tiled field where half the field was subirrigated
(SI) and the other half drained only (DR). Alternative management schemes including
different plant varieties and row spacings were replicated on both the SI and DR plots.
Statistical analysis of yields between SI and DR plots showed no significant difference at
the 5% significance level. Neither was there a yield difference associated with the
different row spacings. There was a slight yield difference between the two varieties. As
was the case with beets and corn in 1989, the drainage only treatment had the highest
economic return because of lower input costs.

Navy beans were gown under the same management systems used in the soybean
trial. Statistical analysis Showed the drainage only yield to be higher than the SI
treatment at the 5% significance level. The economic analysis showed the DR treatment
to be more profitable by $33.85 to $36.53, depending on the row Spacing.

Now turning to the issue of the economic analysis of subirrigation under a
rotation for the same field, available data for yield and net revenue for Site 2 and Site 3
are used to perform a basic economic analysis. For each site, 2 years of data are
available. By taking the figures for the benefit to subirrigation (SI advantage) for each
crop for each year and averaging them, the economic benefit of subirrigation for that
specific rotation can be calculated.

At Site 2, the SI advantage for corn in 1988 was $90/acre and for soybeans in
1989 was negative $18.25 / acre (the average of all SI advantage figures in Table 2.4).
This translates into an average annual per acre benefit to subirrigation for a corn -

soybean rotation of $36 / acre.

42
At Site 3, the SI advantage for sugar beets in 1988 for the 30-ft tile Spacing was

$158. In 1989, the average SI advantage for corn in the 30-ft tile spacing field was
negative $10. This translates into an average annual per acre benefit to subirrigation for

a sugar beet - corn rotation of $74.

1990 Field Trial Results

The 1990 field trials included sugar beet and corn trials at Site 2 and a navy bean
trial at Site 3 (LeCureux, 1991d).

The management scheme for the sugar beet trial consisted of two treatments, an
SI treatment and 8 DR treatment, on a field with 25-ft tile spacings. Both the SI and
DR plots received the same levels of inputs. Rainfall was above the average for the
season.

As can be seen in Table 2.5, the SI treatment yielded 1.8 more tons than the DR
treatment and had a net revenue advantage over the DR treatment of $47.93 /acre.

The corn trial followed the same management procedure as that used in the 1989
beet trial at Site 2. The 30-ft spacing field was subirrigated and the 60-ft spacing field
had both a SI and DR plot. Two varieties of corn were planted to evaluate yield
differences. All plots received the same level of fertilizer, seed, herbicides and
insecticides. The yield results for the two varieties of corn varied only slightly.
Highlighting the results for only one of the corn varieties (Pioneer 3573), the results
show the 60-ft SI treatment had a yield of 151.6 bu/acre and a goss margin of 150.29
while the 60-ft DR treatment had a yield of 142.3 bu/acre and a goss margin of $155.67.

The higher SI yield was not high enough to offset the higher production and fixed costs

43

associated with subirrigation, so the DR treatment had a goss margin $5.38/acre higher
than the 81 treatment in 1990.

The procedure for the navy bean trial was the same used on the corn trial at Site
3 in 1989: 30-ft and 60-ft tile spacings with both an SI and DR treatment on each
spacing. Each plot was planted with three different varieties to determine if there were
substantial yield differences due to plant variety. Rainfall was above average for the
season and only a limited amount of water needed to be applied through subirrigation to
the SI field.

The SI navy beans, regardless of variety or tile Spacing, yielded better than the
DR navy beans. The largest yield difference was 7.3 cwt/ acre for the 30-ft spacing with
the Wesland variety. This treatment also had the largest difference in goss margin,
$107/acre, over the DR treatment. This result is the opposite of what we saw with corn,
where under good rainfall, yield differences between SI and DR treatments were
minimal. It implies that navy beans may benefit more from subirrigation in general,
showing less variability in response due to weather than corn does. This highlights an
important fact: different crops respond differently to subirrigation under weather
variability. Economic analyses of subirrigation under rotation should capture this effect.
With this in mind, the economic analysis of subirrigation under rotation is extended by
incorporating the 1990 yield and net revenue results for the same sites as above.

Using the same approach applied above, the results are as follows:

At Site 2, the 1988-89-90 corn - soybean - sugar beet rotation had net returns of
+$90/acre, -$18/ acre, and +$48/acre, for an average annual per-acre benefit to

subirrigation of $40.

44
At Site 3, the 1988-89-90 sugar beet - corn - navy bean rotation on the 30-ft

spacing field had net returns of +$158/acre, -$10/acre, and +$77/acre, for an average

annual per-acre benefit to subirrigation of $75.

TABLE 2.5: Summary of Field Trials and Economic Analyses for 199010

 

 

 

 

E
Site Crop Mgt Drain Yield Actual SI Net SI
5 ace Goal Yield AOL Rev11 Adv
2 SBT SI 25 28.0 1.8 $929 $48
(SB) DR 26.2 $881
2 CN #1 SI 60 180 151.6 9.3 $150 ($6)
(SBT) 30 180 153.4 5154
DR 60 180 142.3 $156
CN #2 SI 60 180 160.6 8.1 $167 ($7)
(SET) 30 180 162.6 $171
DR 60 180 152.5 H $174 1
3 NB #1 SI 60 25.7 W 5.9 $315 $87
(CN) 30 25.8 4.3 $316 $63
DR 60 19.8 $228
30 21.5 $253
NB #2 SI 60 24.4 3.8 $296 $56
(CN) 30 25.4 4.2 $310 $62
DR 60 20.6 $240
30 21.2 $248
NB #3 SI 60 21.5 6.1 $253 $89
(CN) 30 21.3 7.3 $250 5107
DR 60 15.4 $164
30 14.0 $143

 

 

 

 

 

 

 

 

 

 

 

 

(Source: LeCureux, 1991b)

2.4.2. Simulation Studies

Engineers at North Carolina State University have produced a number of
analyses of drainage and subirrigation using a simulation model, DRAINMOD, that has
been specifically designed to choose the optimum drain spacing in designing a
drainage/subirrigation system (Skaggs, 1981; Hardjoamidjojo and Skaggs, 1982;

Hardjoamidjojo et al., 1982, Skaggs et al., 1982; Evans et al., 1988; Murugaboopathi et

 

’° Seasonal rainfall in 1990 at Site 2 was 18.55 inches and at Site 3 it was 20.70 inches.

“ Corn price = $2.10/bu. Sugar Beet price = $42.00/ton. Navy Bean price = $15.00/cwt.

45
al., 1991). Some of the analyses have focused on the technical aspects of drain spacing

and depth for optimum yields (Skaggs et al., 1982; Hardjoamidjojo and Skaggs, 1982)
and some have looked at the economic tradeoff between increased cost of reduced
spacing between drains and increased yields for drainage and/ or subirrigation (Skaggs
and Nassehzadeh-Tabrizi, 1983; Evans et al., 1988; Murugaboopathi et al., 1991). The
North Carolina studies distinguish between yield benefit due to drainage and that due to
a combination of drainage and subirrigation. They will be presented below based on that

distinction.

2.4.2.1. Benefit to Drainage

Skaggs and Nassehzadeh-Tabrizi (1983) analyzed optimum drainage using
DRAINMOD to simulate corn yields for a 26-year period of North Carolina weather.
They used the results of the simulation model in an economic analysis of the effect of
drain spacing and surface drainage on long-term average profits for corn for two soil
types.

Simulation results showed substantial beneficial effects of subsurface drainage in
wet years and more limited beneficial effects in dry years. Delay in planting date alone
in one particularly wet year would have resulted in a reduction in yield to 65% of the
potential yield even if soil water stresses did not occur during the rest of the gowing
season.

The maximum average predicted relative yield was 78% of potential and occurred
for a drain spacing of 66 feet for good surface drainage and 56 feet for poor surface

drainage. At a drain spacing of 328 feet, which is the conventional Spacing between

46

drainage ditches in North Carolina, the maximum average predicted relative yield was
only 52% of potential (with good surface drainage).

Annual yield results showed that the benefits of drainage are widely variable on a
year-to-year basis. As revealed in the Boggess and Amerling (1983) study, it is the
particular pattern of this variability that affects the expected net present value. This is
not captured in an economic analysis that looks at average yield differences.

Skaggs and Nassehzadeh-Tabrizi tested the sensitivity of their results to weather
at a particular location by running the simulations for the same soil and drainage system
inputs but with 15 years of weather data from 6 other weather stations in North
Carolina. Their results held for the other weather data and the maximum predicted
relative yield occurred for drain spacings between 49 and 66 feet in all cases.

The yield and drain spacing results from the simulations were used in the
economic analysis to determine the effect of drain spacing and surface drainage on long-
term average profits. Net return to land and management for alternative drainage
treatments was calculated as average annual goss income minus costs. Costs included
annual drainage system costs, which were the initial system costs amortized over
estimated useful lifetime and the variable system costs, and corn annual production costs,
which included both fixed and variable costs.

The economic results Show that average profit is not maximized at the same
drain spacing that maximizes yields because of the trade off between increased cost and
increased yield of reduced drain spacing. Maximum yield for the poor surface drainage
case occurred at a drain spacing of 56 feet, whereas maximum net return was obtained
with a drain spacing of 79 feet. With improved surface drainage, the same relationship

held, although net returns were maximized at a lower level than for the poor surface

47 ‘
drainage case because improved surface drainage costs were relatively high. Skaggs and
Nassehzadeh-Tabrizi tested the sensitivity of the results of the economic analysis to
several factors. In their analysis, the initial drainage system costs were amortized at an
interest rate of 12%. Alternate rates of 10% and 14% were tested in a sensitivity
analysis. The results show that while net returns vary depending on the interest rate, the
drain spacing required to maximize profit remains the same. Sensitivity analysis of the
results to the price of corn also showed that the drain spacings required to maximize net
profits for low corn prices were only slightly larger than for the high prices.

The authors also tested the sensitivity of their results to changes in the assumed
potential yield of 175 bu/ acre, which could be too high or too low for some soils
depending on soil fertility, management practices, and weed and insect problems. The
authors found that the optimal spacings are not sensitive to the potential yield for the
range of conditions considered.

Thus, while interest rate, corn price and assumed potential yield have large
effects on net return, they have only small effects on the drainage design required to
maximize net returns.

Skaggs and Nassehzadeh-Tabrizi (1983) also considered the influence of drainage
on the year-to-year variation in net return. For the period 1950 to 1975, net return was
positive in 21 out of 26 years for good drainage as compared to a positive net return in
only 11 out of 26 years for poor drainage.

The authors extended this basic analysis of variability to calculate the payoff
period for an investment in improved drainage. They stressed that a farmer’s ability and
or willingness to invest in drainage depends more on the length of time required for the

investment to pay for itself. This depends on the size of the initial investment and the

48

increase in profits due to drainage. Using the optimum drain spacing that maximized
average profits, they demonstrated that if all profits were used to pay off the initial
investment, the drainage system would pay for itself in only three crop years.

Another North Carolina study of yield benefit to drainage only (Hardjoamidjojo
and Skaggs, 1982) showed that with the correct drain spacing for drainage only, yields
can reach 80% of their potential. Higher yields (up to 20% higher yields) can only be

achieved by reducing deficient soil water stresses through subirrigation.

2.4.2.2. Benefit to Drainage/Subirrigation

Because studies have shown that improved drainage alone can dramatically
improve corn yields (Hardjoamidjojo and Skaggs, 1982), a clear distinction must be made
between benefit to improved drainage and to subirrigation. Evans et al. made this
distinction very clear in their analysis of controlled drainage and subirrigation systems
(Evans et al., 1988). They used the Simulation model DRAINMOD to analyze the effect
of different drain spacings on yield for a drainage only base case, highlighting the
tradeoff of increased yield from reduced drain spacing and increased tile costs of
reducing the drain spacing. For the drainage only base case, they controlled for cost
differences based on surface drainage characteristies, providing alternate calculations for
both a good and poor surface drainage alternative.

After establishing the base yield and net return figures for improved drainage
under good and poor surface drainage conditions, they again used DRAINMOD to
analyze yield increases due to controlled drainage, subirrigation/drainage, and center
pivot sprinkler irrigation. (Only the subirrigation results will be presented here.) Their

cost calculations for subirrigation took into consideration fixed and variable costs for

49

pump and control structures for two alternate water supplies: a deep well and surface
water. Fixed costs included depreciation, interest, property taxes, and insurance.
Variable costs included repairs and maintenance, fuel, and labor. Their pumping cost
calculations for subirrigation reflected the fact that subirrigation is only 75% efficient.
25% of the water pumped is lost through seepage to nonirrigated areas, thus pumping
charges were adjusted accordingly.

Production costs were also broken down into fixed and variable costs. Production
costs for subirrigation were adjusted to reflect increased nitrogen and harvesting costs
associated with a higher yield goal. In this specific example, a yield goal of 130 bu/ acre
was established for the drainage only simulations and 160 bu/ acre for the subirrigation
simulations. (The implications of setting alternative yield goals with DRAINMOD will
be discussed at length in the methodology chapter).

The simulation results for the alternative scenarios of drainage only and
subsurface irrigation / drainage clearly Show the tradeoff between increased yields with
closer drain spacings and increased system costs. For clarity’s sake, only comparisons of
the fair surface drainage alternative for each scenario are discussed here (Table 2.6).
Results of the good surface drainage alternative are similar.

For the drainage only base case, the highest yield (168.5 bu/ac) was achieved
with a drain spacing of 50 ft; whereas the highest net return per acre was for a drain
spacing of 75 ft. For the subsurface drainage/subirrigation scenario, the highest yield
was for a drain spacing of 33 ft. The highest net return per acre for a well water source
was $136.56 at a drain spacing of 50 ft and the highest net return per acre for a surface

water source $164.33 at a drain spacing of 50 ft.

50

TABLE 2.6: Yield and Net Return Results with Fair Surface Drainage

 

 

 

 

 

 

 

 

=1
I ' Optimum Associated Optimum .Associated
Predicted Tile Net Return Yld / Tile
Yield Spacing per acre Spacing
Scenario (bu/acre) (ft) ($[ac) (yld/sp)
Drainage Only 135.6 50 $135.25 134.6/75
Subirrigation
Well 168.5 33 $136.56 162.9/50
Surface 168.5 33 $164.33 162.9/50
ourco: vans ot a ., )

 

 

To establish the additional benefit to subirrigation, the highest predicted net
return for subirrigation was compared with the highest predicted net return for the
drainage only case. Using this criterion, the benefit of subirrigation over drainage only
depends on the water source. Whereas subirrigation is only marginally more profitable
than drainage when a deep well is used as the water source, it can boost average profits
by $29.08 per acre when a surface water supply is available. These results Show that the
cost of the water source can be an important factor affecting profits with subirrigation.

The authors conclude that for the conditions assumed, subirrigation would be the
most profitable choice. But they also raise the point that since the net profit with
subirrigation is only slightly higher than that with conventional subsurface drainage, some
farmers might not want to take the risk of the additional capital outlay.

To address this issue of risk, Evans et al. (1988) considered year-to-year variation
in profit over a 10-year period for alternative drain spacings in a conventional drainage
system compared to several alternative subirrigation systems which varied by drain
spacing and water source (surface water versus gound water). After determining the
tile spacing that gave the highest long-term average profit for each option, the authors
found that conventional subsurface drainage provided the most profit for a continuous

corn production system, but it also had the highest loss in one year out of the ten.

51

Subirrigation provided the most consistent year-to-year profit. A net profit was
predicted every year. The authors concluded that from the standpoint of stabilized farm
income, subirrigation might be the most desirable option, but from the standpoint of
long-term average profit, subsurface drainage might be optimal.

Murugaboopathi et al. (1991) extended the above analysis to evaluate the
sensitivity of results to soil type and to evaluate the impact on optimum drain spacing
and on optimum net returns of including soybeans in a corn - soybean rotation. The
economic analysis was conducted to determine the drain Spacing that gives the maximum
net return to land and management for the corn - soybean rotation.

The procedure used mimics the Evans et al. (1988) study in almost all respects,
with the exception that Murugaboopathi et al. did not include a controlled drainage or
sprinkler irrigation scenario, they used a longer period of weather data (37 years), and
they included much more detailed breakdown of the pumping system characteristics
necessary for alternative drain spacings.

As with the previous study, to simplify matters, only the results for the fair
surface drainage alternative for one of the soil types (Table 2.7 and Table 2.8) are
presented here. Results for the good surface drainage case and for the other soil

evaluated are quite similar to those presented below.

52
TABLE 2.7: Results for Corn for a Rains Soil: Fair Surface Drainage

 

 

 

  

 

 

 
    

 

 

 

 

 

  

 

   

 

 
    

fl_ T.

Optimum Associated Optimum .Associated
Predicted Tile Net Yld / Tile
Yld (bu/afife) Spacing Return Spacing

Management System and YR(%) (ft) (S/ac) (yld / sp)

Drainage Only 130.7 50 $70.89 126 / 80
758

Subirrigation 168. 7 33 L 5111 .89 155.2 / 50
96%

(Source: Nuruga oopathi et al., 1531) l
TABLE 2.8: Results for Soybeans for a Rains Soil: Fair Surface Drainage
Optimum Optimum
Predicted Associated Net Associated
Yield Tile Return Yld / Tile
(bu/acre) Spacing per acre Spacing
Scenario and YR(t) (ft) (S/ac) (yld / sp)

1. Drainage Only 61.6 50 $132.22 59.5 / 80
88%

2. Subirrigation 67.2 33 $122.59 61.6 / 66
96%

 

(Source: Muruga oopat 1 et a .,

 

 

 

 

1)

For the drainage only treatment, the maximum average relative corn yield was

75%. Drought stresses prevented the relative yield from reaching 100% of the potential

yield. For the SI treatment, reducing drought stresses allowed corn yields to reach 96%

of potential. The yield results also Show that the maximum relative corn yield with

subirrigation is obtained at a narrower drain Spacing (33 ft) than with drainage only,

where the maximum relative yield is obtained at 50 ft.

For soybeans, there is much less difference between the maximum average

relative yield for SI and DR. With DR, soybean relative yield can reach 88% of

potential, compared to 78% for corn. These results imply that soybeans are much less

responsive to subirrigation than is com.

 

" YR (%) is the relative yield.

 

 

53

The economic analysis revealed an important issue. In looking only at the corn
results, the benefit to subirrigation was $41 / acre. Subirrigation increased profits by
57.7% over drainage only. This is a very impressive gain. But in looking at the soybean
results, subirrigation resulted in a loss of $9.63 / acre compared to drainage only.
Analyzing the corn - soybean rotation, the optimum drain spacings of 66 ft for the Rains
sandy loam and 98 ft for the Portsmouth soil resulted in increased net profits for
subirrigation over conventional drainage of 18% for the Rains soil and 22% for the

Portsmouth soil.

2.5. Directions for the Current Economic Analysis of Subirrigation

The studies summarized above provide valuable backgound. The issues they
have raised, the approaches that were taken in the economic analyses, and the gaps they
left have Shaped the direction of the current study.

The Huron County and other Michigan studies of subirrigation and drainnage
have provided Site specific information about costs, soil types, system design and
management. The North Carolina studies have shown how useful simulation can be in
analyzing an investment decision that has a long horizon. Simulation provides flexibility
to look at water table management from several different angles and draw important
conclusions about system design and the profitability of subirrigation and drainage. The
two general economic studies of irrigation provided added insight into how economists
approach investment analyses from a somewhat different perspective than
noneconomists.

The results of all of the above studies have led to the following delineation of the

decision setting in the current analysis:

54
The analysis is based on a hypothetical farm in Huron County in the Saginaw Bay

area of Michigan. The soil at the farm is a Kilmanagh soil, the prevalent soil type in the
county. The total number of cropped acres is 400, but the investment decision concerns
only a 40-acre field where a drainage system is in place.

Three scenarios are hypothesized concerning water availability:

1) The farmer has access to a private surface water source.

2) The farmer has the potential to exploit gound water resources.

3) The farmer has the opportunity to participate in an irrigation district where a

water use fee of between $25 and $35 /acre must be paid.

The investment options under consideration include modifying the existing
drainage ornly WTMS by reducing the drain spacing to 20 feet (DR20) or 30 feet
(DR30), keeping the existing drainage only system intact (DR60), converting the existing
drainage only WTMS into a subirrigation system at 20-ft tile spacings for a surface water
source (51208) or a well water source (SIZOW), or into a subirrigation system at 30-ft tile
spacings for a surface water source ($1308) or a well water source (SI30W), or into a
subirrigation system at 60-ft tile spacings for a surface water source (SI60S) or a well
water source (SI60W).

Chapter 3 provides a thorough accounting of the cost assumptions used and the

approach taken in the economic analysis.

2.6. Gaps in the Study
Both the Huron County and North Carolina studies stressed that farmers use a

rotation scheme and because the benefits to subirrigation-drainage vary widely by crop, a

55

thorough economic analysis should explicitly consider the benefit of subirrigation to the
rotations commonly used in subirrigated fields. Due to time limitations, the present
economic analysis considers returns to corn alone. Corn was chosen because it is the
largest cash crop gown in Huron County and therefore has a large economic
significance for the county. In addition, of the four crops most often gown in a rotation,
corn, soybeans, dry beans, and sugar beets, it is the one that has the lowest net returns
per unit of output and the economic results could be interpreted as being a lower bound
on what farmers could expect from investing in improving their water table management
system.

A second gap is that no attempt has been made to incorporate either positive or
negative environmental spillover effects in the economic analysis of subirrigation.
Several studies described above in the water quality impact section and water availability
section (Fogiel, 1992; Fogiel and Belcher, 1991; Kittleson et al., 1990a, 1990b; Kittieson
and He,1990; He et al., 1991, 1992) have provided some insights into the environmental
implications, but the present level of scientific knowledge of these effects is sufficiently
limited that it is premature to incorporate environmental impacts in the present analysis.
AS mentioned above, current research by Michigan State University researchers at the
Saginaw Rain Shelter site will eventually provide some of the necessary environmental
data necessary to extend the economic analysis of subirrigation to include effects on

gound and surface water contamination and erosion.

CHAPTER3

METHODOLOGY

3.1. Introduction

Farmers are faced with many types of uncertainty in their decision environment.
Production risks and financial risks are two of the most important sources of uncertainty.
Production risk includes weather variability, the threat of pest infestations and disease,
and uncertain consequences of production management decisions such as the choice of
plant variety and cultural practices or the timing of production activities. Financial risk
includes price risk and uncertainty stemming from the financial structure of the business.

Financial risk from price variability of inputs and outputs arises from forces
outside the control of the farmer. On the other hand, financial risk deriving from the
structure of the business is geatly affected by capital investment decisions and choice of
financing arrangements for those investments, both of which are determined by the
farmer.

Financial decisions which leave the farmer highly leveraged can exacerbate
production and price risks because fluctuations in output or prices interfere with the
farmer’s ability to make regular debt payments. Yet many capital investments are in fact
made to reduce production variability. Installing an irrigation system is an example of
such an investment. Irrigation is characterized as a yield-increasing, risk-reducing

strategy. But these advantages must be compared to large investment costs in the

S6

57

irrigation system and increased production costs and variable irrigation costs associated
with pumping. The investment is risky from a financial perspective, so in one sense, by
installing an irrigation system, the farmer might be trading production risk for financial
risk (Boggess and Amerling, 1983).

This tradeoff is quite important with irrigation investment decisions in humid
regions, where the variability in weather patterns over the economic life of an irrigation
investment has a significant impact on the profitability of the investment (Boggess et al.,
1982 and 1983; Boggess and Amerling, 1983). This dilemma arises because of the
sensitivity of expected net present value (ENPV) to the sequence of net returns flowing
from the investment over the lifetime of the system. If several poor rainfall years follow
installation of an irrigation system, the system begins paying for itself immediately. On
the other hand, if several good rainfall years follow, the system is not contributing to
significantly increased returns yet is costing the farmer in principal and interest payments
on the investment loan, or in the opportunity cost of lost interest if the investment was
purchased with cash.

The objectives of this study are outlined in detail in the introductory chapter and
the decision setting is described in Chapter 2. This chapter provides a description of the
general approach used and the specific methods employed in the economic analysis of
alternative water table management systems (WTMSS).

The analysis proceeds in six stages.

(1) The production and investment costs associated with gowing corn under the

different options are determined.

58

(2) The Simulation model DRAINMOD is used to generate corn yield and irrigation
application data for the drainage-only and subirrigation options over a 33-year period of
historic weather data.

(3) Monte Carlo simulation is applied to generate one hundred 30-year NPVS by
drawing randomly from the yield and irrigation volume data.

(4) The expected NPV (ENPV) and standard deviation of NPV (SDNPV) are compared
across systems for both water sources.

(5) The probability distributions of NPV generated in the Monte Carlo simulation are
compared in two stages:

a) using first and second degee stochastic dominance (FSD and SSD) criteria;

b) using the stochastic dominance with respect to a function efficiency criterion to

compare those distributions which are not stochastically dominated by FSD or

SSD.

(6) A sensitivity analysis is performed to compare the outcome of the base analysis with
outcomes of analyses run with a range of product prices, investment costs, and yield
assumptions.

Both the NPV analysis using the base weather sequence and the Monte Carlo
simulations are performed using an investment analysis computer progam written in
Quick Basic by the author. The progam allows geat flexibility in changing parameter
values of key variables for the sensitivity analysis. A copy of the QuickBasic progam
code is included in Appendix A.

The methods, key assumptions and relevant theoretical backgound for the
simulation, the NPV analysis, the Monte Carlo simulation, and the risk analysis are

described in detail below.

59

3.2. Simulation
3.2.1. Simulation as a Tool in Economic Analyses

Simulation has become an important tool for agicultural economists interested in
studying decision making and risk analysis at the farm level (Anderson, 1974a). Sources
of risk in agicultural production, including pests and weather, are an integal component
of simulation models. In a figurative sense, the simulation models provide researchers
with a computerized experimental plot (Boggess et al., 1983) where they can hold
constant key decision variables in an agicultural production system and observe the
results of different scenarios under a long sequence of historical weather.

For example, when considering several alternative drain spacings in a drainage or
subirrigation system, the effect of a particular drain spacing under drainage only and
subirrigation can be compared for the same weather sequence. Then the drain spacing
can be changed and the same weather sequence run and the results compared for the
different drain spacing and management regimes. This ability to make repeated
comparisons of high relative precision in the same environments is unique to Simulated
experiments (Anderson, 1974a).

Simulation has been extremely useful in economic analyses that consider risk.
Risk analysis requires information on the probability distribution of decision alternatives.
Yet actual data at the farm level are rarely available for long enough time series to
derive these probability distributions. Simulated yield data can fill this gap. For
example, subirrigation has only been actively practiced in Michigan for just over a
decade, yet hourly precipitation data is available Since 1958. In this case, 33 years of
yield data can be simulated for a technology that has been gowing in use only in the last

10 years and for which the availability of actual yield data is very limited.

60

In addition, simulated yield data have an advantage over historic data in that
technology is held constant in the simulation. This obviates the need to detrend the
data, thus making it easier to isolate variation in output due to risk influences.

While allowing much flexibility, simulation also has certain drawbacks. As with
any modeling exercise, Simulation cannot possibly capture all of the complexity of an
agicultural production environment. The ceterus paribus assumption of economics with
all of its drawbacks holds for simulation as well. Most models tend to focus on one
source of variability at a time. For example, weather variability over a long historic
record is captured in the present analysis, but the soil is assumed to be a stable water-
holding matrix of constant fertility and structure and pest and disease influences are
ignored.

Some critics argue that even the variability in weather that is captured by using
historic weather records in simulations Still provides only a restricted sample from a
stochastic process (Anderson, 1974a). In the current analysis, random draws are made
from the “parent" yield distribution with the assumption that the historic weather-yield
distribution provides an acceptable representation of expected future weather-yield
outcomes. Given the lack of accurate weather forecasting, this approach is warranted,
provided that the results are presented with the appropriate caveats about the

assumptions used in the analysis.

3.2.2. DRAINMOD
The simulation model used in this analysis to study the effects of drainage and
subirrigation on yields is DRAINMOD (version 4.01). It was developed by researchers

at North Carolina State University at Raleigh to study and better design multicomponent

61

water management systems comprised of surface and subsurface drainage and / or
subirrigation, and/ or sprinkler irrigation components. DRAINMOD uses recorded
weather data to simulate the performance of a given site specific drainage design over a
long period of climatological record. The model is described in detail elsewhere (Skaggs,
1981) and will be presented only briefly below.

DRAINMOD uses crop, weather, and soil input data to compute the water
balance in the soil profile. It simulates infiltration and runoff processes based on the
specified drainage system design and then computes daily water table depth, depth of dry
zone at the surface, subsurface drainage, runoff, and evapotranspiration (ET).

The water balance results are used in the crop response component of the model.
The model has three separate components that are summarized in a first-order yield
equation:

YR = YR,,,"‘YR,,"YRp
where
YR is the relative yield,
YR. is the relative yield under wet conditions,
YRd is the relative yield under dry conditions,
YRP is the relative yield if planting is late.

All three terms, YR,” YRd, YRP, are expressed in percentage terms and are
initially assumed to be 100%. Then subtractions from YR”, YRd, and YRp are made for
excessively wet conditions, excessively dry conditions, and delays in planting, respectively.

This approach assumes that there are no interactive effects among the three
model components. For example, it assumes that the effect of excessive soil water

conditions is independent of the existence of deficit soil water at another time in the

gowing season.

62
The relative yield output of DRAINMOD must be converted into a predicted

yield in bushels per acre (bu/acre). The conversion is made by multiplying the relative
yield by a maximum potential yield (POTY) figure expressed in bu/ acre. For the Huron
County Kilmanagh soil of this analysis, the representative maximum potential yield for
nonirrigated corn is assumed to be 140 bu/acre and that for irrigated corn to be 180

bu / acre. A sensitivity analysis of the economic results to changing these POTY
assumptions is performed in Chapter 4.

The input data requirements necessary to run DRAINMOD are extensive. They
fall into three major categories: soil-related inputs, crop specific inputs, and drainage
design inputs. Each is described briefly below and some of the most important input
values used in the simulations are included in Table 3.1. A complete listing of
DRAINMOD input values is included in Appendix B.

Crop specific parameters include planting date, seed bed preparation time, length
of gowing season, and effective rooting depth as a function of time. The actual length
of time required for seedbed preparation and planting depends on size of operation,
equipment and labor available, and many other factors. A period of 8 days was chosen
as being typical for a Saginaw Bay area farm of 400 acres where half the acres are
planted to corn and corn is the crop planted first. Values used for the other parameters

are also representative of Saginaw Bay conditions.l

 

’ Typical values for the Saginaw Bay area were chosen after consulting with a Huron County
Encnsion Agent, James LeCureux, and crop and soil scientists (Dr. Maurice Vitosh and Dr. Francis
Pierce at Michigan State University.

63

TABLE 3.1: Summary of DRAINMOD Inputs

 

 

 

 

 

   
   
 

 

 

. a. -
WW
Depth to restrictive layer 125 - 150 cm
Saturated hydraulic conductivity
‘ x < 70 cm 3.30 cm/hr
70 cm <- x < 112 cm 2.80 cm/hr
I 112 cm <- x < 125 cm < 0.15 gm/qf
: Plant-available water content at wilting 0.22 cm [cm
Required drainage volume for field work 3.40 cm
. Minimum daily rain to stop field work 1.27 cm
, Time after rain before work can resume 2 days
WW
Drain depth 102 cm
Drain diameter 4 inches
Surface depressional storage
poor to fair surface drainage 2.5 cm
tDrain spacings 20, 30, 60 ft
o r e
Crop Continuous corn
Desired planting date Not > May 10
Working time for seedbed preparation 8 days
Length of growing season 105 days
Maximum effective rooting depth 45 cm
Dry slope coefficient 1.05
Net slope coefficient 0.68
SF 1.25
Maximum potential yield
irrigated 180 bu/ac
nonirrigated 140 bu/ac
(A apte rom Evans et al., 1558)

 

Drainage system parameters are drain depth and spacing, effective depth of
impermeable layer, depth of surface depressional storage, drainage coefficient, geometric
parameters used in computing the drainage rate under ponded surface conditions, and
depth of water in the outlet as a function of time. These values are Site and system
specific. A hypothetical field site was chosen and several WTMS designs elaborated by
Dr. Harold Belcher in the Department of Agicultural Engineering at MSU (Appendices
C and D). Values for drain depth and depth of the impermeable layer are
representative of values found in Huron County. They were chosen after consulting
Huron County Soil Surveys (USDA, 1980) and other publications which provide Huron

County specific information (LeCureux and Booms, 1990a-d; LeCureux 1991a,b). A

64

large surface depressional storage parameter was used to reflect poor surface drainage
conditions at the hypothetical site.

Climatological inputs include hourly rainfall and daily maximum and minimum
temperatures. There is no weather reporting station in Huron County that had a long
enough record of hourly precipitation so data was used from the Flint National
Climatological Station for the period 1958 to 1990. Flint is the closest station to the
Saginaw Bay area for which hourly rainfall data are available.2 One drawback of
DRAINMOD is that it requires hourly precipitation data and these data are not readily
available for many stations.

As a check of the similarity of Flint average climatological data and that for
stations in Huron County, a comparison was made between average monthly
temperature and precipitation data for the Flint weather station (Genesee County) and
two stations in Huron County, Bad Axe and Harbor Beach (Appendix E) for the period
1951-1980. Even between the two stations in Huron County, there is quite a difference
in average monthly precipitation, with Harbor Beach receiving more rainfall, especially
during the growing season. Average maximum and minimum temperatures for the two
Huron County sites also differ significantly. Growing season temperatures in Bad Axe
are generally higher than those in Harbor Beach. Comparing the two Huron County
stations with Flint, Bad Axe and Flint have very similar temperatures, while Harbor
Beach and Flint have similar growing season rainfall. Flint tends to have higher average
monthly rainfall than either Huron County station, except during July, when both stations

have higher precipitation than Flint. Bad Axe receives 1.69 inches less rainfall than Flint

 

3 Hourly precipitation data were obtained from the Midwest Regional Climate Center,
Champaign, Illinois. Daily maximum and minimum temperature data were obtained from the
Michigan Department of Agriculture, Environmental Division, Department of Climatology.

65

during the growing season (May-September) while Harbor beach receives only 0.36
inches less than Flint over the growing season. These data are similar enough that the
weather data for Flint could feasibly represent growing conditions in Huron County for
the purpose of this study.

Soil property inputs for DRAINMOD include saturated hydraulic conductivity of
each soil horizon, soil water characteristics, relationships for the drainage volume and
steady upward flux as functions of water table depth, Green-Ampt infiltration
parameters, and water content at the wilting point.

Additional soil-related inputs that allow the model to simulate trafficability
constraints include threshold values for the drainage volume required for field operations
during planting and harvesting periods and for the amount of rainfall necessary to
postpone field operations.

Site specific soil measurements at field sites in Bannister for a Ziegenfuss soil
and Bad Axe for a Kilmanagh soil were provided by Dr. James Crum of the Crop and
Soil Sciences Department at Michigan State University. The saturated hydraulic
conductivity values were taken from Huron County Soil Surveys (USDA, 1980). The
surveys provide a range of possible values for each soil layer. For the base analysis, an
average value was chosen for each layer.

All of the soil properties are very site specific, varying somewhat even within a
given soil type. This necessitates site specific measurements of hydraulic conductivity,
depth to restrictive layer, and soil water characteristics. This is an obvious limitation,
since results of the simulations and the economic analyses derived from the yield results

of the simulations must also be presented as being site specific. However, this is quite

66

representative of reality. Each farm is unique. The present analysis provides results that

farm decision makers must adapt to their own particular situation.

3.2.2.1. Model Validation

DRAINMOD has been validated using field data for North Carolina, Iowa, Ohio,
and India (Skaggs et al., 1981; Hardjoamidjojo et al., 1982; Skaggs et al., 1982). Belcher
validated the water balance component of DRAINMOD using field data from Bannister,
Michigan, for a Ziegenfuss soil (Appendix F).

Validation of the yield component was done for the present analysis using both
aggregate county-level yield data and farm-level yield data. The long-term simulations
for a Kilmanagh soil and the Flint climatological station data were run using the input
data shown in Table 3.1 above. Complete results of these runs are presented in Chapter
4. The results of the drainage-only simulations at a 60-ft spacing (DR60) were validated
against historic Genesee County aggregate corn yield data (Michigan Department of
Agriculture, 1958-91) and historic farm-level corn yield data from a farm near Reese in
Tuscola County.3 The Genesee County aggregate corn yield data were chosen over
Huron County aggregate yield data because the Flint reporting station is in Genesee
County. The Reese farm was chosen because of its relative proximity to the Flint
weather reporting station‘ and because it has a wet loamy soil in the same soil
classification category as the Kilmanagh soil of the analysis. No site-specific weather

data was available for the Reese farm, so the comparability of the simulation predicted

 

3 The Reese corn yield data were made available by Dr. Roy Black of the Department of
Agricultural Economics at Michigan State University from TELFARM records.

‘ Reese is approximately 20 miles directly north of the Flint reporting station.

67

yields and the actual farm-level yields is only approximate. The same is true for the
comparisons of the county-level yield data and the simulation predicted yields. The
comparisons give a global view of closeness of predicted and actual yields in the study
area.

In order to compare DM predicted yields with the historic yield data, it was
necessary to detrend the historic yields to remove the effect of technological change. A
preliminary step before detrending of the data was to test for heteroskedasticity. This
was done by running a regression on corn yield versus time’ (Figures 3.1. and 3.2.) and

plotting the residuals as a function of time.

 

Plot of Genesee Historic Yields
and Estimated Yield with Residual Plot

120

 

100‘

 

Corn Yield (bu/acre)

 

 

 

 

Time

 

+ Genesee Yld -°- Estimated Yld + Residuals

 

 

 

 

 

 

Figure 3.1.

 

5 Yield data for Genesee county are for 1958 to 1991 and for the Reese farm are for 1963 to
1988.

68

 

 

 

 

 

 

 

 

 

 

 

Plot of Reese Farm Historic Yields
and Estimated Yield with Residual Plot
140 _
.20. w A
$1004 A ‘ A A V
3 90‘ v V V ' '
E ,0. U"
V 40-
3 2° A
c 0 K A Ry 7i
3,, \HJ V V \\.
40-
‘601953 19155'1937'19B9'1971j1973H975 1977 1979'19'31'19'33'19'35'19‘37'
Time
+ Reese Yld -l— Estlmated Yld + Residuals
Figure 3.2.

 

Visual inspection of the residuals implied that heteroskedasticity was not present.
However, a second check for heteroskedasticity, Spearman’s rank correlation test was
also preformed (Gujarati, 1978). For the Genesee historical yield data, the computed t
value was 1.47 and for the Reese farm yield data it was 0.79. For both cases the t value
was smaller than the critical t value (t = 2.04) at the 5% significance level, and thus we
can reject the null hypothesis of heteroskedasticity.

The historic and detrended yield data along with the DRAINMOD predicted
yields are presented in Tables 3.2 and 3.3. In these tables, the last column, DIF, shows
the difference between the detrended historic yields and DM predicted yields (DIF =
historic detrended yields minus DM predicted yields). For the Genesee County data, the

average difl'erence is -0.7 bu/acre with a standard deviation of 13.2 bu /acre. For the

 

69

Reese farm data, the average difference is -0.1 bu/ ac with a standard deviation of 22.4
bu/ac. Figures 3.3. and 3.4. show graphically the detrended historic yields and the
DRAINMOD predicted yields.

For the aggregate Genesee county yields, DRAINMOD accurately predicted
most downward trends in yield and only once predicted a much lower yield when in fact
the yield was high (1974). To some extent, in the case of low yield predictions when
yields were in fact low, DRAINMOD tended to exaggerate yield losses (for example, in
1963, 1965, and 1978). For the aggregate county-level data, the standard deviation of the
predicted yields was higher than for the detrended historic yields. Aggregation of yields
to the county level tends to mask variability at the farm level (Fulton et al., 1988), so the
higher standard deviation of the predicted yields is valid since the simulation is field
specific.

For the Reese farm data, DRAIN MOD performw less well. As expected, the
detrended farm-level yields show much more variability. The standard deviation is 21.4
bu/ acre compared with 11.4 bu /acre for the county-level data. DRAIN MOD predicted
yields for the 1963-1988 period had a standard deviation of 19.1 bu/acre, which was only
slightly lower than that for the farm-level yields, but it missed some important downward
trends in yield, missing eight of twelve significantly low historic yields. In addition, twice
it predicted low yields when in fact yields were high. The weather data which the Reese
farm experienced could be significantly different from those recorded at the Flint
reporting station, so this comparison is a rough one. For both the county-level and farm-
level yields, for example, DRAINMOD predicted very low yields for 1963 when the
historic data shows average yields. However, inspection of the Flint weather for 1963

shows a period of 33 days in a critical growth stage where no significant rainfall fell.

TABLE 3.2: Genesee County Historic and Detrended Yield Compared with DM Yields

7O

 

 

 

 

 

 

 

 

 

DM Pre-

Hietoric Detrended dicted

Yield Yield Yield
Year (bu/ac) (bu/ac) (bu/ac)‘ DIP
1958 55.5 105.8 110.0 -4.2
1959 53.5 102.3 90.4 -11.9
1960 47.4 94.7 94.9 -0.3
1961 58.2 103.9 107.7 -3.8
1962 54.6 98.8 110.0 -1l.2
1963 47.2 89.9 61.2 28.7
1964 60.4 101.6 110.0 -8.4
1965 55.0 94.6 79.6 15.0
1966 69.4 107.5 104.8 2.7
1967 76.0 112.6 109.6 3.0
1968 83.8 118.9 110.0 8.9
1969 70.4 103.9 105.5 -1.6
1970 74.7 106.7 110.0 -3.3
1971 59.0 89.5 110.0 -20.5
1972 75.0 104.0 110.0 -6.0
1973 70.0 97.4 110.0 -l2.6
1974 75.0 100.9 83.7 17.2
1975 56.0 80.4 110.0 -29.6
1976 56.1 79.0 107.8 -29.8
1977 73.1 94.4 110.0 -15.6
1978 73.3 93.1 69.4 23.7
1979 98.8 117.1 110.0 7.1
1980 96.6 113.4 110.0 3.4
1981 100.0 115.2 110.0 5.2
1982 103.8 117.5 110.0 7.5
1983 97.9 110.1 109.9 0.2
1984 91.4 102.1 99.2 2.8
1985 105.4 114.5 110.0 4.5
1986 96.7 104.3 110.0 -5.7
1987 84.0 90.1 73.9 16.2
1988 65.7 70.3 67.7 2.6
1989 94.3 97.3 110.0 -12.7
1990 109.9 111.4 110.0 1.4

1991 101.2 101.2

Avg Dif 101.3 101.4 -0.1
SD Dif 11.4 14.9 13.2

 

 

° Relative Yields were converted to bu/ac yields by multiplying by a potential yield of 110 bu/ ac.

 

71
TABLE 3.3: Reese Farm Historic and Detrended Yield Compared with DM Yields

 

 

 

 

 

DM Pre-

Historic Detrended dicted

Year Yield Yield Yield

(bu/ac) (bu/ac) (bu/ac)7 DIP
79.0 125.4 72.3 53.1
58.0 102.6 130.0 -27.4
60.0 102.8 94.1 8.7
65.0 106.0 123.9 -17.9
98.5 137.8 129.5 8.3
55.0 92.5 130.0 -37.5
82.7 118.4 124.7 -6.3
118.3 152.2 130.0 22.2
63.3 95.4 130.0 -34.6
97.6 127.9 130.0 -2.1
78.0 106.5 130.0 -23.5
104.0 130.8 98.9 31.8
122.2 147.2 130.0 17.2
79.5 102.7 127.4 -24.7
87.3 108.7 130.0 -21.3
75.0 94.6 82.0 12.6
103.0 120.8 130.0 -9.2 a
135.0 151.1 130.0 21.1
115.2 129.5 130.0 -0.5
33.0 145.5 130.0 15.5
121.0 131.7 129.9 1.8
130.0 138.9 117.3 21.7
92.0 99.1 130.0 30.9
130.0 135.4 130.0 5.4
88.0 91.6 87.4 4.2
72.0 73.8 80.0 -6.2 I

118.0 118.7 -0.7
21.4 19.1 22.4

 

 

 

 

 

 

Therefore, it appears that DRAIN MOD accurately predicted a large yield loss under
those conditions and the discrepancy possibly derives from differences in localized
weather patterns. As a final check of the validity of DRAINMOD’s output, the yield and
irrigation application amounts for drainage only and subirrigation were shown to Dr.
Jeffrey Andresen, an agrometeorologist at Michigan Department of Agriculture’s
Climatology Division. He judged them to be reasonable predictions after examining the

daily weather data used in the simulations.

 

7 Relative Yields were converted to bu/ac yields by multiplying by a potential yield of 130 bu/ac.

72

 

120

a
O

‘1
O

 

Comaarison of Genesee Yields
ith DM Predicted Yields (“ 1 10)

 

ml A m A 1

Corn Yleld (bu/ac)
4'
A
-
’\
J
x ‘

 

 

 

60
1958 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988

I I I I I l T I l I T j TTTTTTTTTTTTTTTTTTTTT

Time

 

+ Genesee YLD -+- DMKIZ

 

 

 

 

Figure 33.

 

160

Comparison of Reese Farm Level Yields

With DM YR (x 130) - dr6k12

 

150‘

. 1 AA .
ill N1,” ‘1 x

\

 

 

 

 

1933'19'55 ‘1937 r19397971 '19'73‘19'75 '1977'1979 '19'97'19'93 '19?!» Win

Time

 

+ Reese Yld -+— OH Rel Yld ’ 130

 

 

 

 

Figure 3.4.

 

 

73
The validation indicated that DRAINMOD provides a sufficiently reasonable

approximation to the yield response to weather variability to justify using its results for

the economic analysis.

3.3. Economic Analysis

In the economic analysis, base values of financial parameters are used. The
values were chosen as representative of actual conditions in the Saginaw Bay area to
represent a base scenario. In order to determine the sensitivity of the results of the
economic analysis to changes in parameters such as marginal tax rate, price of output,
the discount rate, etc. a sensitivity analysis is performed as a final component of the

economic analysis.

3.3.1. Base Calculations

Annual gross income for each WTMS is calculated from DRAINMOD’s
predicted relative yield (YR), the assumed potential yield (POTY), and the price of corn
(PC) as follows:

Annual Gross Income 2 YR / 100 * POTY * PC.

The potential yield is defined as the yield that would be obtained if soil water
conditions were ideal during the entire growing season. The potential yield for
subirrigated corn is taken to be 180 bu /acre while that for nonirrigated corn is taken to
be 140 bu/acre. These figures were chosen based on historical records for the Saginaw
Bay area (Michigan Department of Agriculture, 1958-1991) and after discussion with an
extension agent in Huron County, James LeCureux, who is familiar with subirrigated and

conventional drainage production systems.

74

The base corn price is held constant at $240/bu throughout the analysis. Since
price risk is constant across management strategies in any particular year. Sensitivity
analysis is carried out for alternative output prices as described in Section 3.4.

These calculations represent the annual gross income for a given water table
management system, up; the increase in gross income over that which would be obtained
for the existing system, DR60. In Chapter 4, separate calculations of gross margins over
the existing system are made in both the base weather sequence NPV analysis and in the
expected value-variance analysis.

System costs include both the investment costs and variable costs associated with
the operation of the system. The initial investment costs for the various WTMS options
vary depending on the water supply source and the drain spacing. In this analysis, it is
assumed that a conventional drainage system with 60-ft tile spacings is already in place
on a 40-acre field (Appendix C). The investment options include:

DR20: Modify the existing drainage system by adding 2 extra laterals between the

existing laterals so that the resulting drainage system has 20-ft tile spacings.

DR30: Modify the existing drainage system by adding 1 extra lateral between the

existing laterals so that the resulting drainage system has 30-ft tile spacings.

DR60: Maintain the existing system at 60-ft tile spacings.

SIZOS: Retrofit the existing drainage system to a subirrigation system with tile

spacings at 20 feet and a surface water source for irrigation.

81308: Retrofit the existing drainage system to a subirrigation system with tile

spacings at 30 feet and a surface water source for irrigation.

SI6OS: Retrofit the existing drainage system to a subirrigation system with tile

spacings at 60 feet and a surface water source for irrigation.

75

SIZOW: Retrofit the existing drainage system to a subirrigation system with tile
spacings at 20 feet and a well as the water source for irrigation.
SIBOW: Retrofit the existing drainage system to a subirrigation system with tile
spacings at 30 feet and a well as the water source for irrigation.
SI60W: Retrofit the existing drainage system to a subirrigation system with tile

spacings at 60 feet and a well as the water source for irrigation.

TABLE 3.4: Description of Investment Options

 

 

 

 

 

 

 

 

 

      
     

 

 

 

Investment Option Water Source Description
DR20 None Drainage only at 20-ft tile spacing
DR30 None Drainage only at 30-ft tile spacing
DR60 None Drainage only at 60-ft tile spacing
(existing system-mo investment)
SIZOS Surface Subirrigation at 20-ft tile spacing
81308 Surface Subirrigation at 30-ft tile spacing
SI6OS Surface Subirrigation at 60-ft tile spacing
SIZOW Well Subirrigation at 20-ft tile spacing
SI30W Well Subirrigation at 30-ft tile spacing
SI60W Well Subirrigation at 60-ft tile spacing

 

      

Site specific designs for each of nine different possible WTMSS were sent to six
drainage/subirrigation contractors. Four of the six contractors provided cost estimates.
In general the subirrigation/ drainage contractors do not install water supply systems so
estimates for the well and the pump were obtained from pump supply firms and well

drilling firms.

76

The cost estimates provided by the drainage/subirrigation contractors for
retrofitting the existing drainage system to either a drainage-only system at narrower
drain spacings or a subirrigation system at three alternate drain spacings all fell within a
range of $1,000. The cost estimates for the well and the pump varied significantly
enough that sensitivity analysis of the economic results to these cost estimates is
performed in the economic sensitivity analysis section of Chapter 4. Table 3.5 contains a
summary of the average costs of individual components of the various WTMSS and Table
3.6 presents the total investment and annualized per-acre investment cost for each
alternative based on a 40-acre system. Appendix G contains the complete cost estimates
provided by the drainage/subirrigation contractors. For the economic analysis, the initial
investment cost is broken down into per acre figures. The investment costs for the
existing conventional drainage system are set to zero.

For depreciation calculations, the life expectancy of each of the different system

components is as follows:

Control structure : 30 yrs
Irrigation risers : 30 yrs
Deep well : 30 yrs
Pump and electric power unit : 15 yrs
Drainage tile : 30 yrs

There are no salvage values anticipated for any of the system components.

Calculation of depreciation was done based on the straight line method. The
pump is depreciated over 7 years and the other system components over 15 years. The
short-term interest rate is 10.5% and the after-tax required real rate of return is assumed

to be 4%.

77

TABLE 3.5: Summary of Component Costs for a WTMS

 

 

 

 

Initial
Component Description and Specifications Cost
Drainage tubing 4 inch corrugated plastic pipe $0.37/ft
6 inch diameter water supply pipe $3.17/ft
6 inch main $1.18/ft
8 inch main $1.63/ft
10 inch main $2.66/ft
12 inch main $3.49/ft
(costs include installation)
Water Supply
nggp_ggll 8-inch, gravel-packed, 100 ft deep, $15,000
10-ft vertical lift, 200 gal/min
SI pump a 7.5 hp pump and electric motor $ 2,000
power unit Installation (includes intake and $ 3,000
discharge lines)
Electrical Service Hookup $ 400
Sggggge watg; River, Stream, Creek, Lake, Canal
SI pump 8 3.5 horsepower pump rated at 200 $1,200
power unit gal/min
Installation (includes intake and $3,000
discharge lines)
Electrical Service Hookup $ 400
Control Head Stands 5 626
‘ Structure Irrigation Inlets $ 117

     

Sourcesta

 

(includes installation costs)

 

 

as"rept¢befitat ve an- contractor estimates.

 

78

TABLE 3.6: Total Investment and Annualized Per-Acre Investment Cost for a 40-Acre
System

 

       
 

 
 
    
    
  
 
  

  
   

   

 

 

 

 

 

 

 

 

Annualized
Drain Total Investment
Water Spacing Investment Costs
Source H __ _ (per acre)
None 20 5 21,062 $ 30
None 30 $ 10,748 $ 16
None 60 Base Case Base Case
Surface 20 $ 37,740 $ 55
81308 Surface 30 $ 27,475 $ 40
' SIGOS Surface 60 $ 17,381 $ 25
Well 20 $ 53,540 $ 77
Well 30 5 43,275 $ 63
Well 60 S 33,181 $ 48
- I I I l

 

 

 

 

 

 

 

 

Water table management system operating costs can be broken down into labor
costs, electricity costs, and repairs and maintenance costs for the different system
components.

Conventional drainage systems do not require management. Subirrigation
systems do. Different management tasks include removing flashboards from the control
structure during wet periods, replacing these boards after sufficient drainage has
occurred, and monitoring the water table level in the field. It is assumed that these tasks
require one-quarter hour per day during the irrigation season and labor is valued at
$6.00/hr for the cost calculations.

Electricity costs depend on the number of acre-inches of irrigation water applied
annually and the per acre-inch cost of pumping. The per-acre annual irrigation
application amounts are one of the outputs of DRAINMOD and the per-acre-inch

pumping costs vary by water source. Based on Huron County studies (LeCureux and

79

Booms, 1990a-d; LeCureux, 1991a,b) average pumping costs were set at $2.25 per acre-
inch for a well water supply and at $1.50 per acre-inch for a surface water supply. Total
per-acre annual electricity costs are calculated by multiplying the acre-inch costs by the
number of inches applied during the season.

The operating and repair costs for different system components were determined
from relevant publications (Evans et al., 1988) and from interviews with the Huron
County extension agent, Jim LeCureux. Operating costs for control structures and
irrigation risers are assumed to be 1% of the average annual investment cost of each of
these components. Repair and maintenance costs for pump are assumed to be 5% of
average annual investment cost. All operating costs are divided by 40 acres to convert
them into per acre figures.

The operating costs for each system component are summarized in Table 3.7 and

the total system operating cost for each alternative WTMS is included in Table 3.8.

80

TABLE 3.7: Variable Costs Associated with Water Management Systems

 

Description, specifications, u
Component and bases for cost Cost
calculations

 

Repairs/Maintenance
Irrigation riser 5
control structure

Fixed percentage of average 1%/yr
annual depreciation

 

 

 

 

Well None assumed
II
F Pumps, power units Fixed t of average annual 5%/yr
depreciation
Electricity
SI System
Well 7.5 horsepower pump $2.25/in
Surface source 3.5 horsepower pump $1.501in
Labor
Subirrigation Based on 1/4 h/day from May
system 15 to Aug 15 to check water

level in observation wells,
adjust riser level, etc. at
$6.00/hr, 40 acres $3.40/ac

 

 

 

TABLE 3.8: System Repair and Maintenance Costs

 

 

 

 

 

 

 

 

 

 

Annual System Annual System
Operating Operating Costs
System Costs (per acre)
DR20 $ 0.00 $ 0.00
DR30 S 0.00 $ 0.00
DR60 $ 0.00 S 0.00
SIZOS $ 34.00 S 0.85
81308 $ 34.00 s 0.85
81608 5 34.00 $ 0.85
SIZOW $ 40.00 $ 1.00
SIBOW 5 40.00 S 1.00
SIGOW 5 40.00 $ 1.00

 

 

 

 

 

81

Enterprise budgets developed by the MSU Department of Agicultural
Economics were consulted for production cost data (Nott et al., 1992) and adjusted
based on Huron County specific production cost data (LeCureux and Booms 1990a-d;

LeCureux, 1991a,b). The values used in the analysis are summarized in Table 3.9.

TABLE 3.9: Irrigated and Nonirrigated Corn Production Costs

 
   
       
   
    
   

Expense

   

 

Seed

Nitrogen 28.60 22.10
% Phosphate 15.20 7.60
Potash 20.90 15.40
Insecticide 12.56 12.65
Equipment Repairs 18.00 18.00

Building Repairs 3.00 3.00
Total:

  

 

 

 

 

It should be noted that production costs do not include depreciation, insurance,
rent, interest, or labor charges. Harvesting costs, which vary depending on the number
of bushels harvested, are calculated separately and include costs for drying fuel, gasoline,
fuel, oil, trucking, and marketing. These costs were summarized in a per bushel
harvesting cost variable (PBC = $0.57) and multiplied by the number of bushels
harvested each year under the various WTMSS to give variable per bushel production

costs (VPBC).

 

' SI = Subirrigated Corn with 180 bu/acre yield goal.

9 N1 = Nonirrigated Corn with 140 bu/acre yield goal.

82
3.3.2. Net Present Value Analysis

3.3.2.1. Theory of Profit Maximization

The base net present value (NPV) analysis looks at the investment decision in
risk-free terms. The base weather-yield sequence is used to derive NPV, which is a
measure of the relative profitability of the different WTMS options. It is assumed that
decision makers’ preferences for NPV can be embodied in a utility function U(NPV) and
that under the conditions of certainty depicted in this first stage of the economic
analysis, the decision maker seeks to maximize utility by maximizing NPV. The decision
choice facing the decision maker is simply to choose the investment with the highest
NPV. In the second part of the analysis, the risk analysis, the assumption of certainty is
dropped and the reality of risk is introduced. Under conditions of risk, the object is to
maximize expected utility. In the results chapter, a comparison of the outcomes of the

two different approaches to analyzing the investment decision is made.

3.3.2.2. Procedures
The procedure used in the NPV analysis is adapted from Boggess and Amerling
(1983). Equation 3 is the formula used for NPV. Each of the variables is described

below and the base values used in the analysis are summarized in Table 3.10.

s .. (1)
NPV = -C.. + 21",- (WC,+VPC+wacp-q-«zvgwmwmcM.__(1”D .2 __D1
r-I (1+k)‘ 1.1 (1 +k)‘

83

where

Co = initial investment cost,

P = price of corn ($240/bu),

Yt = yield in year t,

IVC, = irrigation variable cost in year t,

VPC -= corn production cost,

VPBC, = per bushel harvesting cost in year t,

D, = tax-related depreciation charged against the irrigation system in year t,
i = interest rate charged on operating capital (10.5%),

I = investor’s marginal income tax rate (28%),

k = investor’s after-tax minimum acceptable real rate of return (4%),
n = life of the system in years (30).

Table 3.10: Base Parameter Values

 

 

 

 

IParameter Base Value “
P $2.40/bu
i $0.105
I $0.28
k $0.04
PBCt $0.57/bu
n 30 yrs ”

The first term of Equation 3 is the initial cash outlay. In this analysis, it is
assumed that the farmer pays all of the initial investment costs out of equity.
Discussions with the Huron County extension agent, James LeCureux, revealed that most
farmers pay for their subirrigation systems out of harvest earnings. This assumption
geatly simplifies the analysis because issues such as the farmer’s leverage ratio, loan
payback periods, and long-term interest rates can be set aside. However, including the
discount factor in the analysis accommodates the fact that the cash outlay has an
Opportunity cost associated with it that is captured despite the simplifying assumption of

a cash purchase of the system.

84

The second term is the discounted sum of after-tax income. Depreciation, which
is a deductible expense, is subtracted from goss income. The term (l-L)/(1+ke)‘ is the
tax and discount adjustment factor.

Because NPV is based on after-tax cash flows rather than net income flows and
depreciation is not a cash expense, the income stream must be adjusted by adding
depreciation expenses back into the analysis to prevent double counting them in the cash
outflows. The final term of Equation 3 reflects this adjustment (Boggess and Amerling,
1983).

All net cash flows are expressed in constant 1992 dollars. The discount rate is
also a real rate of return. A real interest charge on operating capital is included in the
calculations. This includes interest on variable production costs, variable irrigation costs,
and variable per bushel harvesting costs.

Because different components of the WTMS have varying life expectancies, the
NPV formulation in the Basic computer progam is actually a variant of the above
formulation. The documented source code in Appendix A provides full details of how
the NPV calculations accommodated this complication.

The output of the first stage of the economic analysis is the NPV of the
alternative WTMS under the base weather sequence. The base NPV results provide a

risk-neutral ordering of the systems.

3.3.3. Monte Carlo Simulation
Monte Carlo simulation is used to address the importance of the sequence of
weather on the profitability of water table management investments in humid climates.

Monte Carlo simulation involves using random numbers in sampling from a particular

85

distribution (Rubenstein, 1981). In this analysis, the process simulation model
DRAINMOD provides a 33-year sequence of yields and irrigation application amounts
derived from running the simulation for 33 years of climatological data. This 33-year
sequence is an historical empirical distribution. By randomly drawing 30-year sequences
with replacement from this historic distribution, we can capture the significance of the
sequencing of weather on the NPV of the investment.

The sequence of Monte Carlo simulation can be depicted as follows:

1. DRAINMOD is used to generate yields (and irrigation application amounts

for subirrigated treatments) for 33 years of daily historical weather data at a

drain spacing of 60-ft.

2. A particular yield response (and irrigation application amount for subirrigated

treatments) is selected by randomly drawing an observation from the uniform

distribution of simulated results.

3. After-tax cash flow for the year is computed using the selected yield and

irrigation application amount in tandem with system-specific costs.

4. Steps 2 and 3 are repeated 30 times. At the end of 30 simulated years, the

net present value of the water table management investment is computed.

5. Steps 2, 3, and 4 are repeated 100 times to generate the probability

distribution of the net present value of the system.

6. Steps 1-5 are repeated for each combination of drain spacing, water supply

source, and water table management option.

(Modified from Boggess, et al., 1983, p. 87)

86

The output of the Monte Carlo simulation for the alternative water table
management investment options provides us with the probability distributions that are

used in the expected value variance and stochastic dominance analyses.

3.3.4. Risk Analysis
3.3.4.1. Risk Efficiency Models

Much of decision theory under uncertainty is based upon the expected utility
model (EUM) which relies an expected utility maximization as its choice criterion. The
utility function embodies information about the decision maker’s preferences. It relates
the possible outcomes of a choice to a single-valued index of desirability (King and
Robison, 1984). It is thus an exact representation of preferences and therefore has much
intellectual appeal. The model has limited practical application, however, because of the
difficulties of estimating utility functions.

One way of getting around this problem is to use an efficiency criterion to order
choices. After specifying certain restricting assumptions about a decision maker’s
preferences, an efficiency criterion divides the decision alternatives into an efficient set
and an inefficient set. The efficient set of alternatives contains the preferred choice of
any member of the class of decision makers for whom the criterion applies (King and
Robison, 1981a).

The benefit of using an efficiency criterion is that by keeping the restrictions on
the utility function rather general, only limited information about preferences is needed
and the efficient set conforms to the Utility functions of a broad class of decision makers.
A disadvantage, however, is that if the restrictions are kept too general, not many

choices will be eliminated as inefficient. As more restrictions are put on the utility

87

function, this narrows the relevant class of decision makers to whom the efficient set
applies and increases the discriminating power of the efficiency criteria. But more
restrictions imply more knowledge of preferences. The tradeoff of generality versus
discriminating power of the efficiency criteria confronts every analyst who tries to decide
on the appropriate efficiency criteria. The choice in the end depends on the specific
problem to be addressed and ultimately on the amount of information available about
the preference function(s) of the decision maker(s).

There are several widely used risk efficiency criteria. The following four will be
described below:
(1) first degee stochastic dominance;
(2) second degee stochastic dominance;
(3) expected value-variance efficiency;
(4) stochastic dominance with respect to a function, also known as generalized stochastic
dominance.
Stochastic Dominance

Stochastic dominance of one distribution over another is determined by
comparing cumulative distribution functions (CDF) of alternative choices. The CDFS

are the integals over the probability density functions (PDFs) of the random variable, x.

For example, if the PDF is f(x) or Fo(x)‘°, the CDF is defined as follows:

R
PM) = 1 f(x) dx (2)

 

‘° Each successive integation of a PDF is denoted by higher subscripts. For example if Fo(x) =
f(x) denotes the PDF, F,(x) is the integal of Fo(x), F,(x) is the integal of F1 (x), etc.

88

In this formulation, it is assumed that x lies within the interval [a, b] and varies
continuously over this range so that the PDFs are continuous (Anderson, 1974b).

Figures 3.5. and 3.6. show gaphically the PDF and the CDF.

 

Probablllty Density Functlon

 

 

 

 

 

 

 

 

F0
f(x)
x
Figure 3.5.
Cumulative Distrlbution Function

1.0

F1
0

x

 

 

 

Figure 3.6.

89
First Degree Stochastic Dominance (FSD)

The concept of stochastic dominance rests on broad assumptions concerning the
preferences of the decision maker. First degee stochastic dominance (FSD) is based on
the assumption that more is preferred to less, i.e. that U,(x)ll > 0. If comparison of
two CDFs shows that one is clearly less than the other, i.e., that F1(R) < = Gl(R) for all
R in [a, b] with strict inequality for at least one value of R, the distribution f(x) is said to
dominate g(x) by first-degee stochastic dominance. Graphically, this means that the
CDF of the dominant distribution can never lie above the CDF of the dominated
distribution (Figure 3.7.). FSD implies that E(U,) > E(Us), which, in turn, means that
f(x) is preferred to g(x).

Thus, without knowing anything more about the utility function other than that
U1(x) > 0, we can say that decision makers with such a utility function will prefer an
FSD distribution (Anderson, 1974b). If one distribution does not dominate the other by
FSD, the two alternatives are both considered efficient by the FSD criterion.

Following this logic, a series of pairwise comparisons is made of the various
alternatives. The comparison can be made gaphically, as described above, where any
CDF which lies entirely above a second is considered dominated by the second. By
eliminating all alternatives that are dominated, an efficient set of choices is thus
determined for the finite set of alternatives under consideration (King and Robison,
1981a).

In order to further reduce the number of alternatives, second degee stochastic

dominance criteria can be applied to the alternatives in the efficient set.

 

" U,(x) is the first derivative of the utility function, U(x). Higher order derivatives are shown
using successively larger subscripts.

90
Second Degree Stochastic Dominance (SSD)

Second degee stochastic dominance (SSD) criteria are used in cases where one
distribution does not clearly dominate the other by FSD, i.e., where CDFs intersect
(Figure 3.7.). SSD criteria are based on the further assumption of diminishing marginal
utility function -- that successive amounts of x have diminishing value to the decision
maker -- U2(x) S 0. Both assumptions taken together, U,(x) > 0 and U2(x) S 0, imply
a concave utility function, U(x). Individuals with a concave utility function are said to be
risk averse.

The ordering rule for SSD is that the distribution h(x) dominates g(x) by SSD if,

and only if,

iH,(R)dRs}G,(R)dR (3)

for all possible R in the interval with strict inequality for at least one value of R
(Anderson, 1974b; Hadar and Russell, 1969). Graphically, this implies that h(x)
dominates g(x) by SSD if area A is not less than area B (Figure 3.7.) or if the
accumulated area under H1(R) is always less than or equal to the accumulated area
under G,(R).

Application of the SSD criterion to a set of alternatives proceeds in the same
manner as for FSD. Pairwise comparisons are made of alternatives. The differences
between the two cumulative probability distributions are summed cumulatively in
ascending order. If the cumulative sum ever changes sign, the pair cannot be ordered by
SSD. If the sign never changes, the alternative with the lower bound of its CDF initially

to the right of the other is the dominant alternative.

91

 

 

 

 

 

 

 

 

Cumulative Distribution Functions
F50 and $30
1.0
cor
o
x
Figure 3.7.

Expected ValueVariance (EV) Efficiency

Expected value-variance (EV) efficiency is also a widely used efficiency criterion.
It assumes (1) risk aversion on the part of the decision maker and either (2a) that the
outcome distributions are normal, or (2b) that the decision maker has a quadratic utility
function. When either 2a or 2b holds, ”all relevant information concerning distributions
of alternative choices is conveyed by means and variances" (King and Robison, 1984,
p.73). If the distributions are normal, EV efficiency criterion is just a special case of
SSD (King and Robison, 1981a).

The ordering rule for EV efficiency is as follows:
f(x) dominates g(x) if { E[f(x)] > = E[g(x)] and Var[f(x)] < = Var[g(x)] } and if at least
one of the inequalities is strict.

The EV efficiency criterion has several advantages over FSD and SSD:

(i) Means and variances are easily derived.

(ii) Most analysts are familiar with the approach.
(iii) It is easily incorporated into quadratic progamming.

92

On the other hand, the EV efficiency criterion shares some of the same
disadvantages of FSD and SSD. The assumption of risk aversion means that for some
decision makers who are not everywhere risk averse, a preferred choice may be
eliminated from the efficient set. In addition, EV efficiency often does not effectively
reduce the choice set.

EV efficiency tends to be inferior to FSD and SSD in at least one respect. The
normality assumption of the EV criteria is often violated in agicultural settings.
Empirical evidence indicates that agicultural yields and other measures of returns have
negative skewness (Day, 1965). In Chapter 4, probability distributions of the expected
NPV of the various WTMS are shown gaphically and it will be seen that the
distributions do show negative skewness. According to King and Robison (1984), if the
normality assumption is violated, the EV efficient set can differ from the SSD efficient
set. For this reason, both EV efficiency criteria and stochastic dominance criteria are

applied to the economic results to determine if they identify the same efficient set.

Stochastic Dominance With Respect to a Function (SDRF)

The efficiency criteria discussed so far all suffer from two deficiencies. None will
reliably reduce a large number of choices to a small efficient set that the decision maker
can order directly and each relies on the assumption of risk aversion (King and Robison,
1984; Harris and Mapp, 1986). Stochastic dominance with respect to a function (SDRF)
overcomes these limitations but requires more knowledge of the utility function. SDRF
imposes linnited restrictions on the utility function (King and Robison, 1981b; Meyer,

1977). It orders uncertain choices for decision makers whose absolute risk aversion

93

functions are within specified lower and upper bounds of the absolute risk aversion
function:

we = -U.(x) / U.(x).
where U(x) is a van Neumann-Morgenstern utility function. Ul(x) is assumed to be
positive (more of the good is preferred to less), so a positive value of R.(x) implies a
negative value of U2(x), which in turn implies a concave utility function and hence risk
aversion on the part of the decision maker. .A negative value of R,(x) implies risk loving
on the part of the decision maker.

The solution procedure for SDRF relies an optimal control techniques. The

object is to identify a utility function Uo(x) which minimizes

l
f [G(x) — F(x)] U,(x) dx (4)
0

subject to the constraint that R,(x) lies everywhere between lower and upper bounds
r,(x) and r2(x), i.e., where r,(x) < = R.(x) < =r2(x).

If the minimized outcome of Equation 4 is positive, the CDF F(x) is preferred to
G(x) by all individuals whose risk aversion function lies within the specified bounds
(King and Robison, 1981a). If it is zero, the two alternatives cannot be ordered. If it is
negative, the positions of F(x) and G(x) in Equation 4 must be reversed and the
equation again minimized subject to the same constraint to determine if G(x) is
preferred to F(x).

This ”preference interval” approach (Cochran and Raskin, 1988) requires that the
class of utility functions be explicitly defined, but it still permits avoidance of the

necessity of representing preferences exactly.

94
FSD and SSD can be related to SDRF by specifying the limits on r1(x) and r2(x)

as follows:
For FSD, g(x) = - infinity and r2(x) = + infinity

For SSD, g(x) = 0 and r2(x) = + infinity.

3.3.4.2. Application of Stochastic Dominance Criteria to Water Table Management
Investment Decisions

In this analysis, 30-year NPV from a 40-acre WTMS investment is the random
variable of interest. First and second degee stochastic dominance criteria and stochastic
dominance with respect to a function are used to identify the risk efficient strategies. In
addition, EV analysis is also performed in order to compare the efficient set identified
by the two methods.

Implementation of the SD rules involves pairwise comparison of distributions to
identify and eliminate distributions that are dominated. The practicalities of this
approach are discussed below.

Steps of Stochastic Analysis

1. For each distribution, rank all the values taken by x (NPV) in ascending order.
2. For each distribution, attribute fix) to each xi. (For 100 NPVS, each NPV has a
probability of .01 associated with it.)

3. Graph the CDF of the NPVs of each WTMS.

4. Make pairwise comparisons among distributions by applying F SD criteria to
determine if one distribution dominates the other by FSD.

5. If this is not the case, apply the SSD criterion.

6. If one distribution does not dominate the other by SSD, apply the SDRF criterion.

95
Determination of FSD is done gaphically in this analysis. Due to limitations in

gaphing capability, CDFS are not displayed in the traditional fashion (Figure 3.8.).
Rather than having the cumulative probability on the y-axis and the expected NPV on
the x-axis, the axes had to be reversed in order to display more than one CDF on each
gaph. Under the circumstances, dominant distributions lie above dominated
distributions, i.e., f(x) dominates h(x). Any distributions that cross cannot be ordered by

FSD and must be evaluated using SSD criteria, i.e., h(x) and g(x).

 

 

 

 

 

 

 

 

 

 

 

 

 

Cumulatlve DIStrIbutlons
with the X-Axis and Y-Axis Reversed
1800
1700*
% 16004
>
E 1500‘
3 1 '
2 1400‘i
n. 172-7
g 1300 ~ ..
120011
1100 'Vl'!I'VVIIlYIUVTIJYJlITII"!"I"V'V'Y'VITIITIV'T'II'lll‘iuuVllllTIT"IIII‘IIT‘[I'V'VVI'V'VIVITI
0.01 0.11 0.21 0.31 0.41 0.51 0.61 0.71 0.81 0.91
Cumulative Probability
—o-— H“) + F(x) -6— G(x)
Figure 3.8.

FSD eliminates from the FSD efficient set any distributions that lie entirely
below a given distribution. Determination of SSD is done by numerically integating the
cumulative difference between distributions which intersect and therefore cannot be

ordered by FSD. If the cumulative sum of the difference between two such distributions

96

does not change sign at any point, the distribution that begins above but subsequently
crosses and remains below the other is said to dominate the one below it by SSD.

A computer progam for ordering distributions using SDRF criteria, GSD version
3.0 (Cochran and Raskin, 1988), was used to evaluate distributions which were not

dominated by either first or second degee stochastic dominance.

3.3.4.3. Application of Expected Net Present Value Criteria to Water Table
Management System Investment Decisions

Expected NPV and standard deviations of NPV are compared across systems. In
any case where the ENPV of one WTMS is geater than or equal to that of the other
and the WTMS also has a standard deviation lower than or equal to the other, with at
least one of the inequalities being strict, it is said to dominate the other by the EV

criterion.

3.4. Sensitivity Analysis
The results of the economic analysis depend on the values of parameters chosen
in the base scenario. In order to test the sensitivity of the results to the particular values
chosen, sensitivity analyses are performed on the following parameters:
a) yield,
b) marginal tax rate,
c) product price,
d) discount rate,

e) initial cost of the WTMS.

97

The sensitivity analysis consists of substituting different values for each of the
above parameters and determining whether the basic stochastic dominance relationships
among WTMSS changes significantly. In some instances, comparisons of changes in the
ENPV and SDNPV under the base scenario and the adjusted scenario are made, but the
emphasis is generally on noting how changes in key parameters affect the stochastic

dominance ordering of the investment options.

CHAPTER 4

RESULTS

4.1. Simulation Results

Simulations were conducted for 33 years (1958-1990) using climatological data
from Flint, Michigan, for alternative drain spacings of 20, 30, and 60 feet. Both
subirrigation and conventional drainage were simulated for each drain spacing and
complete simulation results for each system are presented in Appendix H. Tables 4.1
and 4.2, which show the results for conventional drainage and subirrigation at 30-ft tile
spacings, are presented in the text below for easy reference. Values in the Predicted
Yield column are the overall relative yield multiplied by the potential yield, which is 180

bu/acre for subirrigation and 140 bu/ acre for drainage only.

4.1.1. Drainage Only Results

Referring to Table 4.1 and Appendices H1-H3, the simulation results show that
at the 20- and 30-ft tile spacing for drainage only, excess water stress does not result in
yield reductions nor does wet stress ever result in a delay in planting at these spacings.
'This result is a function of the narrow drain spacing which provides excellent drainage.
As the tile spacing increases to 60 feet some slight excess water stress occurs during 3 of
the 33 years of the simulation. Although simulation results with tile spacings geater

than 60 feet are not included in the appendix, simulations were run at 70-, 80-, 90-, and

98

99
100-ft tile spacings and excess water stress became more severe at successively wider tile
spacings.

The most significant yield reductions for the drainage only scenario at 20-, 30-,
and 60-ft tile spacings occur in all cases as a result of drought stress, which causes yield
reductions geater than 10% in 8 of the 33 years at each tile spacing. As discussed in
the previous chapter, inspection of rainfall data by an agometeorologist at the Michigan
Department of Agiculture Climatology Department, Dr. Jeffrey Andresen, confirms that
the indicated years were in fact relatively dry years based on the daily weather records
from the Flint reporting station.

The simulation results show a predominance of 100% yields. In fact, in just over
half of the years, the predicted relative yield is 100%. Under actual field conditions, one
would assume that the yields would show much more year-to-year variation. However,
no correction is made for this phenomenon. The rationale for not making a correction is
that under field conditions, in a good year, yields often exceed the planned yield goal (as
evidenced in validation Figures 3.3 and 3.4). Yet DRAINMOD does not predict yields
geater than 100% of the assumed potential yield, so to some extent the large number of
100% yields compensates for this deficiency.

In comparing the simulation results for the conventional drainage case at the
three tile spacings, 20-ft, 30-ft, and 60-ft, there is a general tendency for the predicted
relative yields to increase as the drain spacings increase. This occurs in all years except
1969 where, at the 60-ft drain spacing, yields fall slightly because of excess water stress.
This result is sigiificant, showing that on a Kilmanagh soil where only drainage is
practiced, reducing the tile spacing below 60 feet can occasionally result in yield losses

from over drainage. Later, in the economic analysis section, this fact results in the clear

100

domination of the 60-ft tile spacing over the narrower drain spacings for the drainage

only systems.

TABLE 4.1: DRAINMOD Yield Output for Drainage Only at 30-ft Tile Spacings

Water Stress Relative Yields Predicted
Plant Plant —— Yield
Year Excess Def Date Delay Excess Def Delay Overall(bu/acre)

 

 

 

1958 0 O 125 O 100 100 100 100 140.0
1959 O 22.2 125 0 100 76.7 100 76.7 107.4
1960 O 14.8 125 0 100 84.5 100 84.5 118.3
1961 0 2.9 125 0 100 97.0 100 97.0 135.8
1962 0 0 125 O 100 100 100 100 140.0
1963 O 44.8 125 O 100 52.9 100 52.9 74.1
1964 0 0 125 O 100 100 100 100 140.0
1965 0 28.5 125 O 100 70.1 100 70.1 98.1
1966 0 5.7 125 0 100 94.0 100 94.0 131.6
1967 O 0.4 125 O 100 99.6 100 99.6 139.4
1968 0 0 125 0 100 100 100 100 140.0
1969 0 O 125 O 100 100 100 100 140.0
1970 O O 125 O 100 100 100 100 140.0
1971 O 0 125 0 100 100 100 100 140.0
1972 O O 125 O 100 100 100 100 . 140.0
1973 0 O 125 0 100 100 100 100 140.0
1974 2.6 23.7 125 O 98.5 75.1 100 74.0 103.6
1975 O 0 125 0 100 100 100 100 140.0
1976 0 0.4 125 O 100 99.6 100 99.6 139.4
1977 O 0 125 O 100 100 100 100 140.0
1978 O 37.6 125 0 100 60.5 100 60.5 84.7
1979 0 O 125 O 100 100 100 100 140.0
1980 0 0 125 0 100 100 100 100 140.0
1981 O 0.2 125 O 100 99.8 100 99.8 139.7
1982 O 0 125 0 100 100 100 100 140.0
1983 0 1.2 125 O 100 98.8 100 98.8 138.3
1984 O 12.3 125 0 100 87.1 100 87.1 121.9
1985 0 0 125 0 100 100 100 100 140.0
1986 0 0 125 0 100 100 100 100 140.0
1987 0 33.1 125 0 100 65.2 100 65.2 91.3
1988 O 38.2 125 0 100 59. 100 59.9 83.9
1989 O O 125 0 100 100 100 100 140.0
1990 O 0 125 0 100 100 100 100 140.0
Average 0 8.1 125 0 100 91.5 100 91.5 128.1

4.1.2. Subirrigation Results
Referring to Table 4.2 and Appendices H4-H6, the results for the subirrigated
simulation runs show much less variability in yields. This can be expected. Irrigation is

practiced to reduce yield variability. The difference between the drainage only case and

101

the subirrigation case is that the yield losses under subirrigation at the narrower drain
spacings result from excess water stress rather than drought stress. Excess water stress
results under subirrigation because the water table is being held at a high level and when
a rainfall event occurs under these circumstances, water invades the top 30 cm of the soil
causing excess water stress and yield reductions.

DRAINMOD captures this effect very well. By varying the weir elevation in the
control structure in the field, and analogously the weir setting in DRAINMOD, these
yield losses due to excess water stress can be eliminated. However, a conscious decision
was made to leave the weir setting in DRAINMOD at a level (55 cm) that resulted in
some excess water stress because this more closely represents reality. Farmers do not
have total control of the water table depth on a continual basis. Having an optimal weir
level setting in DRAINMOD would imply superior management which is not the case in
field situations.

Consequently, the simulation results show wet stress under subirrigation
dirnirnishing as the drain spacing increases. Conversely, at the 60-ft spacing, drought
stress begins to cause yield reductions. This reflects actual experience in Huron County.
At a wider drain spacing, water does not move laterally through the soil far enough to
reach the middle portion of the field between two tiles. Inspection of the irrigation
volumes (Appendices Il-I6) shows that in moving from 20- to 30- to 60-ft tile spacings,
successively smaller volumes of water can be pumped out through the tiles. At the 60-ft
tile spacing water becomes limiting enough that drought stresses result.

A gaph of the predicted yields converted into bushels per acre is presented in
Figure 4.1 for the drainage only system and the subirrigation system at 30-ft tile spacings.

This gaph clearly shows that because of the difference in the assumed potential

102

maximum yields between the two systems (180 bu /acre for subirrigation versus 140
bu/acre for drainage only), the subirrigated system consistently has higher yields than the
drainage only system. The sensitivity of the economic analysis to these differences in

assumed potential yields is tested below in the economic sensitivity analysis section.

TABLE 4.2: DRAINMOD Yield Output for Subirrigation at 30-ft Tile Spacings

Water Stress Relative Yields Predicted
Plant Plant Yield
Year Excess Def Date Delay Excess Def Delay Overa11(bu/acre)

 

 

 

1958 0 0 125 0 100 100 100 100 180.0
1959 5.9 0 125 0 96.0 100 100 96.0 172.8
1960 0 0 125 0 100 100 100 100 180.0
1961 0 0 125 0 100 100 100 100 179.5
1962 0.5 0 125 0 99.7 100 100 99.7 180.0
1963 0 0 125 0 100 100 100 100 180.0
1964 0 0 125 0 100 100 100 100 180.0
1965 0 0 125 0 100 100 100 100 180.0
1966 0 0 125 0 100 100 100 100 180.0
1967 1.7 0.1 125 0 98.8 99.9 100 98.8 177.8
1968 11.9 0 125 0 91.9 100 100 91.9 165.4
1969 8.3 0 125 0 94.3 100 100 94.3 169.7
1970 7.3 0 125 0 95.0 100 100 95.0 171.0
1971 0 0 125 0 100 100 100 100 180.0
1972 10.2 0 125 0 93.1 100 100 93.1 167.6
1973 0 0 125 0 100 100 100 100 180.0
1974 4.9 0 125 0 96.7 100 100 96.7 174.1
1975 0.2 0 125 0 99.9 100 100 99.9 179.8
1976 2.6 0 125 0 98.2 100 100 98.2 176.8
1977 0 0 125 0 100 100 100 100 180.0
1978 0 0 125 0 100 100 100 100 180.0
1979 0.7 0 125 0 99.5 100 100 99.5 179.1
1980 3.4 0 125 0 97.7 100 100 97.7 175.9
1981 0 0 125 0 100 100 100 100 180.0
1982 0 0 125 0 100 100 100 100 180.
1983 0.7 0 125 0 99.6 100 100 99.6 179.3
1984 0 0 125 0 100 100 100 100 180.
1985 0 0 125 0 100 100 100 100 180.0
1986 0 0 125 0 100 100 100 100 180.0
1987 0 0 125 0 100 100 100 100 180.0
1988 0 0 125 0 100 100 100 100 180.0
1989 0.4 0 125 0 99.7 100 100 99.7 179.5
1990 0 0 125 0 100 100 100 100 180.0
Average 1.8 0 125 0 98.8 100 100 98.8 177.8

103

 

DRAINMOD Predicted Yields

for Drainage Only and Subirrigation

180- v "w v V v w

100-

Yield in bu/acre
3'3
0

80‘

 

 

 

‘0 I I j—I I—fiI I I I I I I I I I If I I I I I I I I I T T I I I I
1959 199011192 11194 1999111991970 1972 197419791979 19901992 199419911 1999 1990
Time

 

+DR Y|d+S|Yld

 

 

 

 

 

 

Figure 4.1

4.2. Results of the Economic Analysis

The economic analysis proceeded in three stages: an analysis of NPV using the
base weather sequence, an analysis of expected NPV, and a risk analysis using the
probability distribution of NPVs which were derived using Monte Carlo simulation

techniques. The results of all three stages are discussed below.

4.2.1. NPV Analysis - Base Weather Sequence
As described in the methodology chapter, the NPV for each WTMS was

calculated using the following formulation of the NPV equation.

104

NPV ' ‘Co + 2.31”;- (IVC,+VPCWPBC)-D,-1'(IVC,+VPC+mung]._(l ’0 .2 __D'
n1

(1+k)‘ r-l (1+k)‘

The results of the economic analysis of NPV under the base weather sequence

(1958-1987) over the 30-yr planning horizon are shown in Table 4.3, which includes a

”Grass Margins” column showing the difference in the NPV of the various WTMS over

the existing system, DR60. In addition, Table 4.4 shows the cumulative NPV over the

entire 30-year planning horizon for each system. These yearly figures show how the final

NPV figure is derived. They were printed out after each loop of the NPV calculation.

TABLE 4.3: NPV and Gross Margins - Base Weather Sequence (1958-87)

 

 

 
     
  
  
  
  
  
  
   
 

 

 

L
Annualized Gross
Net Net Margin Annualized
Investment Present Present Over Gross

Option Value Value DR60 Margins
DR20 $ 954 $ -$ 446
DR30 $1,164 5 -$ 236
DR60 $1,400 3 ------
81208 $1,341 5 -$ 59
81308 $1,598 5 $ 198
51608 $1,761 5 s 367
sxzow $1,019 5 -$ 381
SI30N $ -$ 124
S $44

 

 

 

 

A basic interpretation of the NPV figures can be stated as follows: Under the

base weather sequence, a farmer who has an existing drainage system in place and who

is gowing continuous corn can expect the present value of his/ her net income stream

over a 30-year planning horizon to be $1,400 per acre. Dividing $1,400 by the value

105
17.292 from a Present Value of Annuity Table (Harsh et al., 1981) for 4% real interest

and a 30cyear planning horizon, this figure can be annualized and interpreted as meaning
the farmer would be indifferent between receiving 581 in annual per acre net returns
over the 30-year period and receiving $1,400/ acre today.

In comparing the $1,400 figure to the NPV for the other WTMS options, it is
clear that WTMS options DR20, DR30, 81208, SIZOW, and S130W are not profitable
while WTMS options SI30S, $1608, and SI60W offer the farmer an opportunity to earn
more per acre than he/she can expect to earn with the existing WTMS. These results
indicate that under a no risk situation, if the farmer chooses only to maximize NPV, only
three of the six subirrigation WTMS options are more profitable than the existing
drainage only WTMS.

Annualized goss margins, which are also included in Table 4.3, give a global view
of what the level of annual returns over the returns from the existing system might look
like. 81608 has the largest annualized goss margin of $21/acre. This figure could be
used as a basis to determine a willingness to pay measure if an irrigation district were to
be formed. However the $21/acre figure would have to be considered an upper bound
of what farmers would be willing to pay since in the calculation of production costs for
all WTMSS, labor costs and fixed costs such as insurance, land rent, and any depreciation
and interest costs not associated with the WTMS investment itself were not included.

The cumulative NPV figures in Table 4.4 give a better idea of the ”payoff period"
of each investment alternative under the base weather sequence. Negative figures in
Table 4.4 indicate that the initial cost of the investment has not yet been recuperated.
Generally, in a goss margins type analysis where results are reported in terms of goss

margins over the existing system, the payback period would be considered the period

106

where the NPV stream is negative and then just becomes positive. In this analysis,
although results in Table 4.4 are not reported in terms of goss margins over the existing
system, the point at which the NPV stream becomes positive is still referred to as the
payback period. A separate distinction is made between the payback period, as used in
this context, and the point where the investment under consideration yields a NPV that

surpasses that of the existing system during the investment planning horizon.

TABLE 4.4: NPV of WTMS Options over the Planning Horizon (1958-87)

 

YEAR DR20 DR30 DR60 81205 SIBOS SI6OS SIZOW SIBOW SIGOW

 

1958 -426 -173 91 -802 -550 -303 -1190 -938 -690
1959 -369 -119 150 -680 -427 -175 -1060 -808 -556
1960 -302 -55 212 -549 -301 -53 -922 -675 -427
1961 -217 26 290 -423 -180 65 -790 -547 -302
1962 -131 108 368 -302 -63 177 -663 -424 -184
1963 -115 121 381 -186 49 261 -541 -306 -94
1964 -35 197 453 -76 157 365 -425 -192 16
1965 2 231 486 29 257 457 -315 -86 114
1966 67 294 547 129 354 546 -210 15 207
1967 138 361 610 222 445 635 -112 111 301
1968 206 425 672 288 522 718 -42 192 388
1969 271 488 727 366 600 795 40 274 469
1970 334 548 784 444 676 875 123 354 553
1971 394 606 839 526 755 952 209 438 634
1972 452 661 891 595 823 1025 281 509 712
1973 503 712 942 649 877 1080 325 553 756
1974 528 737 969 715 941 1136 392 617 812
1975 575 783 1016 777 1005 1200 454 682 877
1976 619 828 1059 824 1066 1261 501 743 938
1977 663 872 1102 884 1126 1321 561 803 998
1978 673 882 1115 942 1184 1373 619 861 1051
1979 713 922 1155 989 1239 1429 667 916 1106
1980 752 961 1193 1034 1290 1480 711 968 1158
1981 788 998 1231 1083 1339 1530 761 1017 1207
1982 824 1033 1266 1131 1387 1577 808 1065 1255
1983 857 1067 1300 1175 1432 1623 853 1110 1301
1984 882 1092 1327 1219 1476 1667 897 1154 1345
1985 914 1123 1359 1261 1519 1709 939 1196 1387
1986 944 1154 1389 1302 1559 1750 979 1237 1428
1987 954 1164 1400 1341 1598 1767 1019 1276 1444

 

NPV: 954 1064 1400 1341 1598 1767 1019 1276 1444

 

107
The cumulative NPV results for each goup of innvestments, DR20, DR30, and

DR60 (Drainage Only), 81208, 81308, and SI60S (Subirrigation with a Surface Water
Source) and SIZOW, SIBOW, and SI60W (Subirrigation with a Well Water Source), are
discussed separately below. In each case, the existing WTMS (DR60) is also compared
to the subirrigation options irn each goup to give a better idea of how the existing system
compares with the subirrigation WTMS options for each water source. The cumulative
NPVs for each goup of WTMS are presented gaphically in Figures 4.2 - 4.4. Figures
4.3 and 4.4 show DR60 compared with the subirrigation WTMSS in each goup.

Of the drainage only WTMS (Figure 4.2), the unmodified existing system (DR60)
provides the highest NPV. Because no initial investment is made, returns are positive
over the entire planning horizon, whereas for DR20 and DR30, returns do not become
positive until years 8 (1965) and 4 (1961), respectively.

The simulation yield results already gave us a premonition of this outcome.
Improving drainage by decreasing the drain spacing below the existing 60-ft spacing for a
Kilmanagh soil is not an economically viable decision for farmers. Because of the clear
dominance of DR60 over the other two drainage only WTMSS, in the subsequent
comparisons with subirrigation WTMSS, DR60 is the only drainage only option
considered.

For the surface water subirrigation options, S1605 has the highest NPV of the
three (Figure 4.3). Returns become positive in year four of this investment and the NPV
over the 30-yr planning horizon is $1,767. The payback periods for the other two surface
water subirrigation WTMSs are 8 years for 8120s and 6 years for $1308.

The fact that SI60S is economically more profitable than 31308 is noteworthy.

The yield results for the two systems indicated that S1308 outperformed SI60S because at

108

the 60-ft tile spacing, water could not be pumped adequately to the center of the field
between two tiles. The economic results indicate, however, that the yield benefit of the

narrower tile spacing does not compensate for the extra cost of reducing the tile spacing

to 30 feet.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cumulative NPV for Draina e Only WTMS
Base Weather Sequence (19 8-87)y
1500
1000-1
>
2
O
s): 900-
g
E
o o
«7"!
19'99'19'90'19'92 1994 19'99 '19'99 '19'70 '19'72 '1974 19"79'19'79'19'90'19'92'19'94'19'99
Time
+ DR20 —"" DR30 + DR60
Figure 4.2

In comparing DR60, the existing system, with the the three surface water
subirrigation systems, it is already clear just from looking at the NPVs for these systems
that DR60 is more profitable than 81208 and less profitable than 81308 and 81608. What
is interesting to note, however, is how long into the planning horizon DR60 remains

dominant over 81308 and 81608. The NPV of 81608 does not overtake that of DR60

109
until year 10 of the planning horizon (1967) and the NPV of 81308 overtakes that of
DR60 in year 20 (1977).
The extra information provided by comparing NPV streams of the different
investment Options over the planning horizon is a matter of interest to decision makers.
For example, even given the fact that 81308 eventually provides a higher NPV than

DR60, many farmers would not be willing to wait 20 years for the extra benefit from

their investment to kick in.

 

Cumulative NPV for DR60 and Surface
Water Subirrigation WTMS - Base Weather

2000

 

1500-

1000‘

500‘

 

 

Cumulative NPV

 

 

 

I I I I I I fI T Ifi I I I I I I I I I I T I I I I I I I j
1999 1990 1992 1994 1999 1999 1970 1972 1974 1979 1979 1999 1992 1994 1999
Time

 

+ DR60 + 51205 + $1303 ‘9‘ $1605

 

 

 

 

 

Figure 4.3

The economic results for the WTMS under subirrigation with a well water source

(shown in Figure 4.4) mimic those for a surface water source except that the NPV

 

110

stream does not become positive until later and is lower in each case over the 30-year

planning horizon because of the higher initial investment cost associated with installing a

well. In this case, DR60 has a higher final NPV than either snow or SI30W, and

although SI60W has a higher final NPV than DR60, it only overtakes it in year 27 of the

planning horizon.

 

 

 

 

 

 

 

 

 

 

 

 

Cumulative NPV for DR6 O and Well Water
Subirrigation WTMS - Base Weather
1500 __—..-
1000- r 5 ="‘""'
a 900- ' 5 " "
Z 5 ‘
0 a '4
.2 0
:6 .. '1
3 a '4
g 4004 ,4
Q n
-1000-t
-‘500 I j I I I I I I I I I—I I I I I I I I I I I I I I I I I I I
1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986
Time
—-— 01290 _._ s120w + s130w —a- snow
Figure 4.4

These results occurred because of the particular sequence of weather following

the initial investment. Had more bad weather years followed installation of the

subirrigation systems, the systems would have had a positive NPV stream sooner and

 

 

111

would have dominated the existing system earlier in the planning horizon. This point

brings us neatly to the next stage of the analysis.

4.2.2. Net Present Value and Expected Value-Variance Analysis

Use of Monte Carlo simulation techniques to generate distributions of 100 NPVS
from the alternative investments captures the implications of weather sequence on the
profitability of an irrigation investment. Expected NPV (ENPV) and standard deviation
of NPV (SDNPV) are calculated from the NPV distributions. These results are
presented in Table 4.5. Also included in Table 4.5 is a Grass Margins column which
shows the difference between the ENPV of each WTMS option and DR60, the existing
system. These results are evaluated from two perspectives. First, it is instructive to look
at the ENPV results as if risk were not an issue, i.e., ignoring the standard deviations,
and compare them with the NPV results from above. Second, risk can be addressed by
applying EV efficiency criteria.

Comparing the various options, we see in the Gross Margins column of Table 4.5
that if we ignore differences in standard deviation, DR60 has a higher ENPV than any
option except 81308 and 81608, and therefore would be the preferred choice compared to
the options with lower ENPV under a situation of profit maximization under certainty.
These are the conditions considered above in the NPV analysis for the base weather

sequence. However, the results here are surprisingly different.

112
TABLE 4.5: Expected NPV, SD of NPV, and Gross Margins Over DR60

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Annualized
Gross
S
S
$
5
ISI3OS $1,466 $ 16 s 77 S 4 I
ISIGOS $1,650 8 68 $261 $ 19
* I SI20W $ 782 S 30 -$607 -$ 35
I SI30W $1,073 5 16 -$316 -$ 18
LSIGOW $1,292 5 67 -$ 97 -$ 6

 

 

Table 4.6 reproduces the NPV and ENPV results for easy reference. In general,
the results are similar, as should be expected. However, comparing the two results, the
subirrigation alternatives fared much better under the base weather sequence than under
the Monte Carlo distribution of weather-yield outcomes. The DIF column shows the
difference between the the two results, with DIF = NPV - ENPV. Under the base
weather sequence, the NPV is in each case over $100 geater than the ENPV for the
subirrigation options. This shows that the particular sequence of weather following
installation of the subirrigation system under the base sequence of weather turned out to
be a ”lucky draw“ for subirrigation options. If this had been the only approach taken,
conclusions might have been biased in favor of the subirrigation options. Including the
Monte Carlo simulation, which essentially consists of randomly drawing one hundred 30-

year sequences of weather-yield outcomes to generate a probability distribution of NPVS,

 

’ Gross Margin refers to the difference between the ENPV of the existing system, DR60, and
the ENPV of the investment alternative.

113

allows us to handle the randomness of weather. Comparing the NPV results under the
base weather sequence with the ENPV results gives a much better appreciation of the

sensitivity of NPV results to a particular ”draw" of weather.

TABLE 4.6: Comparison of NPV and ENPV

 

 

 

 

 

 

 

 

 

lsrzos $1,341 $1,187
5130s 91,599 $1,466
I sreos $1,767 $1,650
stzow $1,019 9 792

sraow $1,279 91,073

 

 

 

 

I SIGOW $1,444 51,292
B“

Up to this point, risk has not been taken into consideration. The benefit of

 

subirrigation that is not captured by looking at either NPV or ENPV alone is its
contribution to reducing variability of returns. Inspection of the standard deviations in
Table 4.5 shows that in all cases, the subirrigation options have lower variability of
returns than the drainage only options. Application of EV efficiency criteria to the
drainage only options reveals that the DR60 dominates the other options because it has
both higher ENPV and lower SDNPV than DR20 and DR30. DR60 must be compared
separately with the surface water subirrigation options and the well water subirrigation
options to reflect the two mutually exclusive water source situations available in the

decision environment of this analysis. In comparing DR60 with the surface water

 

3 DIF is the difference between NPV and ENPV, i.e., NPV - ENPV.

114
subirrigation options, it is dominated by 81308 and 81608. Comparison of DR60 with the

well water subirrigation options reveals that the existing system remains in the efficient
set with all the three alternative options.

Using EV efficiency criteria alone, the choice set between 81308 and 81608 and
between DR60 and the well water subirrigation options cannot be further reduced to one
efficient option for each set because the choice depends on the risk preferences of the
decision maker. For example, while 81608 has a higher expected NPV it also shows
more dispersion about that value as measured by a higher standard deviation. 81308 has
a lower ENPV, but it also has a lower standard deviation and might be preferred by
some decision makers who desire more stable returns, even if that means accepting a
lower expected NPV.

Graphically, the variability and level of returns can be easily visualized for the
different WTMS options. The probability distribution of NPV for DR60, SI60W, 81308,
and 81608 is depicted in Figures 4.5 - 4.8. This gaphical presentation gives us a better
appreciation of the tradeoff between variability versus level of ENPV. Comparing 81308
and 81608, we can clearly see that the probability of getting a higher return with 81608 is
quite high, but we can see equally well that a farmer who does not want to risk the slight
probability of the lower returns in the negative tail of the distribution might feasibly
choose 81308, where all the probability is essentially concentrated over the $1400 NPV
interval.

Some of the distributions of NPV are somewhat negatively skewed. This raises
the issue of whether the EV criteria should be used to order distributions because the
normality assumption is violated. Below stochastic dominance criteria are applied to the

different options to see if the same efficient set is identified.

115

 

100

HISTOGRAM OF NPV
r012 DR60

 

00‘

70-1

PROBABIJTY (1N PERCENT)
8

 

 

 

 

 

 

900 ' 700 ' 900 '900 '1000'1100 1200 1900 1400 1900'1900'1700

NPV

 

Figure 4. 5.

 

100

HISTOGRAM OF NPV
r012 5150w

 

90-1
80‘

70-
601

40-

30-1

PROBABILITY (IN PERCENT)
3

20d

10‘

 

O O.
— —

 

 

 

I j I fifn' '." '.I'r 1 I r
000 700 000 000 1000 1'00 1200 1300 I400 1500 I000 I700

N’V

 

Figure 4 . 6.

 

 

 

116

 

HISTOGRAM OF NPV
FOR $1305

100 ”+77

 

 

00-1

41

7°.

00‘

40‘

30‘

PROBABILITY (IN PERCENT)
3

20'

 

10.

 

_
f

0 1900'1900'1700

 

900 '700 '900 '900 '1000'1100'1200'1900'140
NPV

 

 

 

 

Figure 4.7.

 

HISTOGRAM OF NPV
FOR 51603

 

30‘

,,,,,,
9999999

909 so eeeeee

es e‘e' lee
eeeeeeeee
21 es 0,. es

000000000

PROBABILITY (IN PERCENT)
(I
0

use eeeeeeee
O. 0
see 0,0,9 see
so ole, so

101 .........
0. O. 0

e
sssssssssssss
es es

  
   

e e
eeeeeeeeee

900 '700 '900 '900 '1009'1100'1200'1900'1400'1
N’V

 

 

500 1000 1700

 

 

 

Figure 4.8.

117

4.2.3. Stochastic Dominance Analysis

Using a gaphical approach as described in the methodology chapter, application
of FSD criteria allows us to eliminate inefficient distributions. First, DR60 is compared
to the other drainage only options and then to the surface and well water source
subirrigation options. For the drainage only WTMSS, DR60 dominates the other two by
FSD (Figure 4.9). For surface water source subirrigation systems, 81208 is dominated by
81308, 81608, and DR60 by FSD (Figure 4.10). 81608 dominates DR60 by FSD. The
ordering of DR60 and 81308 and the ordering of 81608 and 81308 must be determined
by applying SSD criteria. For the well water source subirrigation systems, DR60
dominates all three options by FSD (Figure 4.11).

In applying SSD criteria, the cumulative difference between the sets of CDFS
(DR60 and 81308; 81608 and 81308) are evaluated to determine whether the cumulative
sum of their differences ever changes sign. In the case of the 81308 - 81608 pair, a sign
change does occur, meaning that the two options cannot be ordered by SSD criteria. For
the DR60 - 81308 pair, no sign change occurs, so 81308 dominates DR60 by SSD.

At this point we are left with a narrower choice set than the EV approach
indicated: 81308 and 81608 still cannot be ordered, but the choice between DR60 and
81308 has been narrowed to 81308 by application of 88D criteria.

As a final step in the risk analysis, the 81608 - 81308 pair is subjected to SDRF

criteria.

118

 

CDFS OF DRAINAGE ONLY WTMSS

 

1900
1500-
1400-
g 1900-
i 1200-
? 1100
i 1000
g 900 '
900
700 -

 

600
0.01 0.11 0.21 0.31 0.41 0.51 0.61 0.71 0.81 0.91
Cumulative Probability

 

+ DR20 + DR30 + DR60

 

 

 

 

 

 

 

Figure 4.9.

 

COMPARISON OF CDFS OF DR60
AND SURFACE WATER SUBIRRIGATION SYSTEMS

 

1800
1700 7
1600 J

 

Net Present Value

 

00
0.01 0.11 0.21 0.31 0.41 0.51 0.61 0.71 0.81 0.91
Cumrlative Probability

 

+ 81205 "+- 51305 4“ $1605 ‘3‘ DR60

 

 

 

 

 

 

 

Figure 4.10.

119

 

COMPARISON OF CDFS OF DR60
AND WELL WATER SUBIRRIGATION SYSTEMS

 

1500
1500 " -:::—::::=:; 33:23:"

......
............

1400« ------- “
1300~ .-

1200-,_

1100 ',

1000 '

900

800

700
0.01 0.11 0.21 0.31 0.41 0.51 0.61 0.71 0.81 0.91

Cumulative Probability

 

Net Present Value

 

 

+ SIZOW -‘— SI30W + SIGOW ‘9'- DR60

 

 

 

 

 

 

Figure 4.11

4.2.4. Stochastic Dominance with Respect to a Function Analysis

In order to discriminate farther between the two top-ranked systems, 81308 and
81608, stochastic dominance with respect to a function was applied using Cochran and
Raskin’s GSD 3.0 progam. The resulting risk aversion interval was adjusted to the scale
of annual income from a 400 acre Huron County corn farm assuming a 4% real discount
rate (following Raskin and Cochran, 1986). As a result, 81308 was found to be
dominated by 81608 for all levels of absolute risk aversion less than .002. This implies
that only a highly risk averse individual would prefer the more costly 81308 system when

com sells for $240/bushel.

120
4.3. Sensitivity Analysis

In the economic analysis, WTMSS have been compared by three methods: Basic
NPV analysis, EV analysis, and stochastic dominance analysis. To simplify discussion in
the sensitivity analysis, emphasis is placed on noting differences in the stochastic
dominance relationship between DR60 and the subirrigation WTMS options under the
base scenario (80) and the adjusted scenarios. In certain cases, reference is made to
changes in the absolute values of expected NPV. Table 4.7 shows the expected NPV and

standard deviation of NPV of the WTMSS under the different sensitivity tests.

4.3.1. Sensitivity to Potential Yield

The first sensitivity test concerned the assumed potential yields. Under the base
analysis, the potential yield under subirrigation (SPOTY) was assumed to be 180 bu / acre
and that for drainage only 140 bu/ acre (DPOTY). Several different combinations of
assumed potential yields were run in the economic analysis to determine the "switching
point,” i.e., the point at which the dominance ordering changed such that DR60, the
existing system, was no longer dominated by any of the other WTMSS as a result of
changing the assumed potential yields. In all cases, when potential yields were changed,
all necessary other changes in variable inputs and associated costs were also made.

In reducing the SPOTY from 180 bu/acre to 170 bu/ acre while holding the
DPOTY at 140 bu/acre (S8), or in keeping SPOTY at 180 bu/acre and increasing
DPOTY to 150 bu/acre (811), the only relationship that changed was between DR60 and
SI30W. Whereas before 81308 dominated DR60 by 88D, under the new yield

relationship between SPOTY and DPOTY, they could no longer be ordered by SSD.

121
In reducing SPOTY to 160 bu/acre but keeping DPOTY at 140 bu/acre (89), the

switching point occurred, meaning that DR60 was no longer dominated by either 81608
or 81308 by 88D criteria.

When the potential yields were brought even closer together so that SPOTY was
set at 160 bu/acre and DPOTY was set at 150 bu/acre (810), DR60 dominated all the
subirrigation options by FSD.

These results indicate that the economic analysis is somewhat robust vis a vis the
assumed differences in potential yields. A 10 bu/acre reduction in SPOTY was
necessary to eliminate the dominance of 81308 over DR60 and a 20 bu / acre reduction in

SPOTY was necessary to eliminate dominance of $1608 over DR60.

TABLE 4.7: Yield Sensitivity Analysis3

   

 

 

 

 

 

 

 

 

 

 

 

Water Table SPOTY 8 170 SPOTY 8 160 l SPOTY 8 180
Management DPOTY = 140 DPOTY 150
Investment “‘_‘“"_““"”“”‘ ” ’"’ "‘ ’ '““ “”

Option ENPV SDNPV ENPV SDNPV ENPV SDNPV ENPV SDNPV
DR20 947 88 947 88 1065 95 1065 95
DR30 1156 88 1156 88 1274 94 1274 94
DR60 1389 82 1389 82 1508 88 1508 88
S1208 1055 29 923 27 923 27 1187 31
$1308 1331 15 1196 14 1196 14 1466 16
81608 1519 64 1389 60 1389 60 1650 68
SI20N 650 28 518 27 518 27 782 30
8130" 938 16 803 15 803 15 1073 16
SI60W 1161 63 1031 60 1031 60 1292 67

 

 

 

 

 

 

 

 

 

 

 

3 ENPV and SDNPV for the base analysis are included in Table 4.8.

 

122
4.3.2. Financial Parameter Sensitivity Analysis

A second set of sensitivity analyses of the economic results to changes in the
assumed tax bracket (TB) and after-tax real rate of return (ATRR) revealed that the
relationship among WTMSs does not change as the tax bracket is either decreased to
15% (85) or increased to 31% (86). However it does change with an increase in ATRR
from 4% to 8% (87). After increasing ATRR, 81608 maintained its position of
dominance over DR60. But DR60 dominated all other subirrigation options by FSD

after the change.

TABLE 4.8: Financial Parameter Sensitivity Analysis

Water Table - Tax Bracket Tax Bracket Base
Management ATTR - 8% TB = 15% TB 2 31% Anal sis

‘ Investment “"""‘_‘—”W""——‘~’— **‘
j Option

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

123
4.3.3. Investment Cost and Output Price Sensitivity Analysis

Changes in the output price significantly changed the relationship among the
WTMSS. At a lower output price of PC = $1.80 (82), which is the price farmers in the
Saginaw Bay area are receiving for corn at the elevators after the 1992 harvest, DR60
dominates all W'I'MS options by FSD. In addition, the expected NPV of DR60 at the
lower output price exceeds that of all other WTMSS. In fact, at the lower output price,
only two of the subirrigation options, 81308 and 81608, had positive ENPV.

At higher output prices this result is reversed. With PC = $2.85/bu (83)
all of the surface water subirrigation options dominate DR60 by either FSD or 88D. For
the well water subirrigation options, 8160W dominates DR60 by FSD and the pair SI3OW
- DR60 cannot be ordered by 88D.

For PC = $3.00/bu (84), which might be considered the upper bound on what
farmers might expect to receive for their corn, DR60 is dominated by FSD or 88D by all
possible subirrigation WTMS options except SHOW.

The minimum corn price at which a subirrigation system stochastically dominates
the alternative of no investment is $2.05/bu under FSD and $2.00/bu under 88D. In

both instances, 81608 is the dominant system.

124
TABLE 4.9: Output Price Sensitivity Analysis

Corn Price Corn Price Corn Price 8389 '
pea 91.90/00 pea 92.95/00 pc= 93.00/00 Anal_sis ;
. ,,____.,L.,_S.______._ _-_______ .___—___—_—_—_—_——_——_.——_—_———‘

ENPV SDNPV ENPV SDNPV ENPV SDNPV ENPV SDNPV

   

I Water Table
Management
. Investment

Option

 
    

 
     
    

 

DR20

 

 

 

 

I
I
|
$1309 139 11 2462 20 2795 22 1466 16“

 

 

 

 

 

 

 

 

 

 

 

 

81608 351 44 2624 85 2948 91 1650 68
SIZOW -528 20 1765 38 2092 41 782 30
SI30W -255 12 2069 20 2402 22 1073 16
SI6OW -7 44 2266 85 2591 91 1292 67

 

 

The final sensitivity analysis involved changing the values of certain costs
associated with installing a subirrigation system. The installation cost estimates were
based on estimates from drainage/subirrigation contractors, pump distributors, and well
drilling firms. The costs given for drilling a well and for buying and installing the
pumping system varied widely, while costs for other system components, including
drainage tile and control structure installation varied only within a small range. To judge
the sensitivity of the economic results to the values used in the base analysis, the cost of
the well and pump were varied separately and then together.

From conversations with drilling firm representatives, the cost of drilling an
irrigation well can vary from $10,000 to $25,000, depending on the specific drilling
conditions. A figure of $15,000 for well drilling was used in the base analysis. Under the
base analysis, all of the three WTMSS with a well water source (8120W, SI30W, and

SI60W) were dominated by DR60 by FSD. If a well drilling cost of $10,000 is used in

125
the economic analysis instead of $15,000 (814), 8160W and DR60 can no longer be

ordered by SSD while DR60 maintains its position vis a vis the other two well water
subirrigation systems, showing that economic results are only slightly sensitive to the
assumed cost of the well in the base analysis.

The pump installation costs used in the base analysis were quoted by contractors,
but appeared to be somewhat inflated, based on figures quoted by the Huron County
extension agent, James LeCureux, who is familiar with pump prices paid by certain
farmers in the county. If the pump installation costs are reduced by 50% and the
economic analysis rerun (813), the results change only slightly. 8160W and DR60 can no
longer be ordered by 88D whereas under the base scenario, DR60 dominated SI60W
FSD. The position of DR60 vis a vis the other WTMSS in the altered analysis remains
the same as in the base analysis.

If both the pump and well costs are reduced together (815), the basic relationship
among the WTMSs changes more noticeably. 81308 dominates DR60 by FSD instead of
88D and SI60W dominates DR60 by FSD, reversing the relationship between these two
options compared with the base analysis. 8130W and DR60 can no longer be ordered by
88D under this scenario.

The result of the cost sensitivity analysis confirm that the economic analysis is
relatively robust to changes in certain key cost parameters. Under all circumstances,
changing the cost of inputs changes the expected NPV of the various options, but rarely

are the relationships among the various WTMSS significantly changed.

126
TABLE 4.10: Investment Cost Sensitivity Analysis

   

   

 

' Water Table
. Management
Investment

  

  
    

 

Pump Cost 8 Well Cost 8
50“ 3°

  

Well 6 Pump

Analysis
SDNPV

   
 
 
 
 

   

   
 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ Option ENPV SDNPV
DR20 947 88 947 88 947 88 947 88'

' DR30 1156 88 1156 88 1156 88 1156 88
DR60 1389 82 1389 82 1389 82 1389 82
$1208 1237 31 1187 31 1237 31 1187 31
S1305 1516 16 1466 16 1516 16 1466 16
81608 1699 68 1650 68 1699 68 1650 68
8120“ 846 30 882 30 947 30 782 30
SI30W 1140 16 1173 16 1238 16 1073 16
SI60N 1356 67 1392 67 1457 67 1292 67

 

 

 

In summary, the ranking of alternative WTMS investment options by stochastic

dominance criteria is most sensitive to changes in yield response, output price, and after-

tax real rate of return. It is less sensitive to financial parameters such as the tax bracket

and the fixed cost of the irrigation pump and well.

CHAPTER 5

SUMMARY AND CONCLUSIONS

5.1. Summary

This study set out to evaluate alternative water table management system
(WTMS) investments. The strategies evaluated include converting an existing drainage-
only system at 60-ft tile spacings into a subirrigation/drainage system at the same 60-ft
spacing (81608 and SI60W), reducing the drain spacing to 20 feet and 30 feet in a
drainage-only system (DR20, DR30), and reducing the drain spacing to 20 feet and 30
feet in a subirrigation/drainage system (81208, 81308, 8120W, 8130W). For each of the
subirrigation options, two different water supplies were considered, a well water supply
and a surface water supply.

The particular strategies chosen for analysis reflect actual conditions in Huron
County and the other counties in the Saginaw Bay area. Many farmers are improving
their existing drainage systems by ”splitting the tiles," i.e., reducing the spacing between
the drains by adding an additional tile line or more between two existing drainage tiles.
Typically, the drain spacing in the area is 60 to 66 feet. Thus farmers are reducing this
spacing, usually to 30 feet. But some farmers who are installing new systems are spacing
the drains at 20 to 25 feet. Reducing the spacing between drainage tiles improves
uniformity of drainage and allows farmers to get onto their fields in the spring for

planting and in the fall for harvest. Other farmers in the area are retrofitting their

127

128

drainage systems from drainage-only systems to subirrigation systems. This study looked
at whether investing in a WTMS is likely to provide enough additional benefit to offset
the cost of the investment.

One other issue of interest was how much financial benefit the dominant surface
water subirrigation strategy might generate toward financing an irrigation district to bring
water from Lake Huron to farmers’ fields. This is an important issue because a large
number of acres in Huron County that are otherwise highly suitable for subirrigation do
not have a sufficient water supply.

A simulation model, DRAINMOD, was used to generate yield and irrigation
application amounts for 33 years of historic weather data from Flint, Michigan.
DRAINMOD was chosen as the simulation model in the yield analysis because it
captures the effect of both excess and deficient water stress on corn yields and it is
specifically designed to study these effects under both subirrigation and drainage at
different drain spacings. For these reasons it was an ideal choice for the present
analysis. However, validation of the yield component of the model proved difficult
because there was insufficient field data at a site where a long enough series of hourly
rainfall data existed. Using the Flint weather data and historic corn yields for Genesee
County, DRAINMOD tracked fairly well the fluctuations in yield. Using Flint weather
and historic farm-level corn yield data, DRAINMOD performed less well. When
DRAINMOD’s output was compared with the daily weather for the Flint station, an
agometeorologist judged the predicted yield results to be realistic.

In the first stage of the economic analysis, DRAINMOD yield and irrigation

application amounts for the base weather sequence (1958-87) were used in conjunction

129

with investment, operating, and production cost data to calculate net present values for a
30-year planning horizon.

Results of the NPV analysis revealed that two of the three surface water
subirrigation options, 81308 and 81608, and one of the well water subirrigation options,
SI60W, had higher NPV than the existing 60-ft drainage-only option. However, the
existing system dominated the remaining subirrigation options and the two narrower
spacing drainage-only options. The annualized goss margins for the dominant
subirrigation systems over the existing system were $11 /acre for 81308, $21/acre for
81608, and $3/acre for SI60W.

The base NPV analysis provided a measure of what a farmer could expect the
NPV of the investment options to look like under the base weather sequence. But it did
not answer the larger question of what a farmer could expect under different weather
sequences. Application of Monte Carlo simulation techniques provided this extra insight.
From the distribution of NPVS generated by drawing randomly one hundred 30-year
sequences of weather-yield outcomes, expected net present values (ENPV) and standard
deviations of net present value (SDNPV) were calculated. The ENPV gives a measure
of how a farmer can expect the NPV of the investment alternatives to look given one
hundred possible 30-year sequences of weather and the SDNPV provides insight into the
variability of ENPV.

Comparison of the NPV and ENPV results showed that under the base weather
sequence, the subirrigation systems fared much better than under the randomized
weather sequences. In all cases the NPV of subirrigation options was more than $100
geater than the ENPV of the same option. Looking only at ENPV and ignoring

SDNPV initially, only the two surface water subirrigation systems at the wider drain

130
spacings, 81308 and 81608, had higher ENPVs than the existing system. 81308 had an

annualized goss margin of $4/acre and 81608 had an annualized goss margin of
$19/acre over the existing system.

For the dominant surface water subirrigation system, 81608, the figure of
$19/acre could be interpreted as the on-farm benefit of subirrigation and could be used
as a measure of the willingness to pay a water use fee in an irrigation district. However,
this figure would have to be considered an upper bound because in the cost calculations,
labor costs and fixed costs such as insurance, land rent, and any depreciation and interest
costs not associated with the WTMS investment itself were not included.

Bringing the SDNPV back into the picture, application of EV efficiency criteria
across investment options revealed that the same two surface water subirrigation
strategies, 81308 and 81608 dominate the conventional 60-ft drainage-only system, DR60,
The dominant strategies have both higher ENPVs and lower SDNPVs than the
conventional system. Between 81308 and 81608 dominance could not be established
under EV criteria. For circumstances where a surface water source is unavailable,
neither could dominance between DR60 and SI60W be established using EV efficiency
criteria because between the two there is a tradeoff between higher ENPV and higher
variability of ENPV. DR60 has a higher ENPV of $1,389 compared with $1,292 for
SI60W, but it also has a higher SDNPV of $82 compared with $67. Thus bringing
standard deviations into the decision framework, 8160W remained in the efficient set
with DR60, as had been the case in the base NPV analysis.

Application of first and second degee stochastic dominance (FSD and 88D)
criteria identified a narrower choice set than the EV approach indicated: 81308 and

81608 still could not be ordered and both still dominated DR60, 81608 by FSD and 81308

131
by 88D criteria. But using SSD criteria, DR60 dominated 8160W, whereas using EV

efficiency criteria, the two could not be ordered.

In order to discriminate between the two top-ranked systems, 81308 and 81608,
stochastic dominance with respect to a function was applied and 81308 was found to be
dominated by 81608 for all levels of absolute risk aversion less than .002, based on
whole-farm annual net income. This implies that only a highly risk averse individual
would prefer the more costly 81308 system.

Sensitivity of the economic results to changes in yield assumptions, output price,
cost assumptions and certain financial parameters was tested by varying these key
parameters. The first sensitivity test concerned the assumed potential yields. Under the
base analysis, the potential yield under subirrigation (SPOTY) was assumed to be 180
bu/acre and that for drainage only 140 bu/acre (DPOTY). A 10 bu/acre difference in
either SPOTY or DPOTY eliminated the dominance of 81308 over DR60 and a 20
bu/acre reduction in SPOTY eliminated dominance of 81608 over DR60. The
dominance ordering between both pairs was reversed completely when SPOTY was set at
160 bu/acre and DPOTY was set at 150 bu/acre. These results indicate that the
economic analysis is only modestly robust vis a vis the assumed differences in potential
yields.

Changes in the assumed output price also significantly changed the stochastic
dominance relationship among the investment options. Lower output prices favored
DR60 and higher output prices favored the subirrigation options. At a price of $1.80/bu
DR60 is no longer dominated by any of the subirrigation options. The minimum corn

price at which a subirrigation system stochastically dominated DR60 was $2.05 /bu under

132
FSD and $2.00 under 88D. In both instances, 81608 was the dominant system. At the

higher prices, all the subirrigation option except 8120W dominated DR60.

Of the financial parameters tested, changing the tax bracket used in the analysis
did not affect the stochastic dominance relationship among the various options; however,
changing the after-tax real rate of return (ATRR) did. Increasing ATRR from 4% to
8% skewed the results in favor of DR60, the existing system. After the change, DR60
dominated all of the subirrigation options by FSD, except 81608, which still dominated
DR60 by FSD. The economic results were robust to changing cost assumptions about

the pump and well.

5.2. Conclusions

This economic analysis of water table management investment options identified
two subirrigation options as dominating the existing drainage-only system (DR60) under
conditions of certainty and of risk. These were both surface water subirrigation systems,
one with tile spacings at 30 feet (81308) and the other at 60 feet (81608). 81608 had an
annualized goss ENPV of $19/acre over DR60. This figure could be used as a measure
of the on-farm benefit of subirrigation for continuous corn production and hence as an
upper bound on farmers’ willingness to pay to obtain a surface water supply (e.g., by
participating in an irrigation district).

In considering the two surface water subirrigation options, a farmer would have
to be extremely risk averse to choose the narrower spaced option, 81308. The 60-ft tile
spacing option, 81608, had an annualized ENPV $15 /acre higher than 81308. Its SDNPV

was also $3/acre higher (in annualized terms); however, the difference in ENPV between

133

the two was larger enough that even moderately risk averse farmers would still choose
81608 over 81308.

None of the well water source subirrigation systems dominated the existing
system and neither did the narrower spaced drainage-only options. These results suggest
that the additional investment costs of drilling a well and the higher pumping costs
associated with deep well pumping offset the benefit of higher and more stable
subirrigated yields and that the additional cost of investing in improved drainage on a
Kilmanagh soil may not produce enough additional yield benefit to offset the investment
costs. These results held under both assumptions of certainty and of risk.

All of the economic findings in this analysis are valid under the assumption that
continuous corn is being produced on the 40 acre field. In the Saginaw Bay area, the
actual practice is to rotate some combination of corn, soybeans, beets, and dry beans.
Including an appropriate rotation in the economic analysis would have to be done to gain
a true appreciation of the economic outcome of investing in a water table management
system. It is a limitation of the current study that these other crops could not be
included due to time constraints.

However, based on the results of other economic studies reviewed here
(LeCureux and Booms, 1990a-d; LeCureux, 1991a,b) it appears that returns to
subirrigation of a rotation including sugar beets would be higher than for a continuous
corn production regime because subirrigated sugar beets produce a substantial net yield
and net revenue benefit over drainage-only sugar beets at recent prices. For the other
two crops commonly in the rotation, soybeans and dry beans, the results are mixed.
Some years they yield a positive net revenue benefit to subirrigation and some years the

benefit is negative (LeCureux and Booms, 1990b; LeCureux, 1991a,b). If we assume

134

their net contribution to the rotation is zero, including a profitable crap like sugar beet
in the economic analysis of subirrigation under rotation would have to increase the net
revenue benefit of the rotation over the continuous corn regime and hence the returns to
subirrigation. Future research will have to look at the broader issue of the profitability
of WTMS investments under a rotation.

The economic results are also sensitive to the assumed corn price. The minimum
corn price at which a subirrigation option, 81608, dominated the existing drainage-only
system was $2.00/bu under 88D. If it were anticipated that corn prices were to remain
below $2.00/bu in the future, farmers should not consider improving their water table
management system. If, on the other hand, it is anticipated that corn prices will be
higher than $2.00/bu, the surface water source subirrigation options at the wider drain
spacings would provide farmers with higher net returns than their existing drainage-only
system if surface water were available for irrigation at no extra cost. If corn prices were
as high as $2.85-$3.00/bu, even the well water options at the 30- and 60-ft tile spacings
would become more profitable than the existing system. Only at a $3.00/bu corn price
would the on-farm benefit of subirrigation at 60-ft tile spacings using surface water
(81608) produce enough additional benefit to offset water use fees as high as $35/ acre.
The necessary corn price to produce an additional on-farm benefit of $25/acre over the
existing system would be $2.70/bu.

Another issue that future research will have to consider is the environmental
spillover effects of alternative water table management investment options. Current

research on these effects should provide the necessary data to conduct such an analysis.

APPENDIX A

BASIC coon FOR MONTE CARLO SIMULATION roe E[NPV] AND sothvn
CALCULATIONS roe WTMS ANALYSIS

DEFINT I-N

'INITIALIZATION OF MODEL PARAMETERS

POUT I 2.4 'corn price in S/bu

TB I .28 'tax bracket

ATRR I .04 'after-tax real return

OCI I .105 'interest on operating capital

NRUNS I 100 'number of simulation runs

NNTMS I 9 'number of WTMS options

NYRS I 33 'number of years of weather data
NNPV I 30 'number of loops for NPV calculation
PBC a .57 'summary variable of all per

'bushel costs, including drying
'cost, harvesting fuel cost,
‘trucking/freight cost, and
'marketing cost.

OPEN "C:\123\DATA\MONTE12A.OUT' FOR OUTPUT AS #1
OPEN "C:\123\DATA\MONTE128.0UT” FOR OUTPUT AS #2
OPEN "C:\123\DATA\MONTE12C.OUT" FOR OUTPUT AS #3

'DIMENSIONING OF ARRAYS

DIM XNPVRUN(NRUNS) 'Array for NPV in the Monte Carlo
'simulation.

DIM D2(NYRS, 2) 'Yield and volume associated with

DIM D3(NYRS, 2) 'different WTMS.

DIM D6(NYRS, 2) 'D2 = drainage: 20-ft tile spacing.

DIM S2(NYRS, 2) '82 I subirrigation 20 ft spacing.

DIM S3(NYRS, 2) 'etc.

DIM serxas, 2)

DIM A(NYRS, 2)

DIM YR(NYRS) 'Array of years 1958-90.

DIM SDNPV(NWTMS), ENPV(NWTMS) 'Arrays used in the SDNPV and ENPV
'calculations.

DIM COUNT(7) 'Array for histograms.

DIM K(NNPV * NRUNS) 'Array for storing the random #3.

'ARRAYS OF COSTS ASSOCIATED WITH THE DIFFERENT INVESTMENT OPTIONS
'SEE BELOW FOR DEFINITIONS. IN ALL CASES, WTMS 3 WATER TABLE MGT SYSTEM

DIM CNTMS(NWTMS), PWAT(NWTMS), VPC(NNTMS), POTY(NWTMS)
DIM DWELL(NWTMS), DPUMP(NWTMS), DTILE(NWTMS), DCS(NWTM8)

'VALUES FOR THE CWTMS I Initial Investment Cost for Each Investment
'Alternative (converted to S/acre figures).

DATA 527, 269, 0, 944, 687, 435, 1339, 1082, 830

135

136

'VALUES FOR PHAT I Price of Water (converted to $/acre-cm)
DATA 0, 0, 0, 0.59, 0.59, 0.59, 0.89, 0.89, 0.89

FOR IA . 1 To Nanns
READ PNAT(IA)
NEXT IA

'VALUES FOR VPC I Variable Production Costs (S/acre)
DATA 105, 105, 105, 128, 128, 128, 128, 128, 128

FOR IA . 1 TO NWTMS
READ VPC(IA)
NExT IA

'VALUES FOR POTY I Potential Yield (bu/acre)
DATA 140, 140, 140, 180, 180, 180, 180, 180, 180

FOR IA - 1 TO NWTMS
READ POTY(IA)
NEXT IA

'VALUES FOR DNELL I Depreciation on the Well for years 1-15 (S/acre)
DATA 0, O, 0' 0' 0' O, 25' 25' 25

FOR IA - 1 TO'NWTNS
READ DwELL(IA)
NEXT IA

'VALUES FOR DPUMP I Pump Depreciation Pump, years 1-7 and 16-22 (S/acre)
DATA 0, 0, 0, 16.43, 16.43, 16.43, 19.28, 19.28, 19.28

FOR IA I 1 TO NWTMS
READ DPUMP(IA)
NEXT IA

'VALUES FOR DTILE I Depreciation on the Tile for years 1-15 (S/acre)
DATA 35, 18, 0, 52, 34, 18, 52, 34, 18

FOR IA I 1 TO NWTMS
READ DTILE(IA)
NEXT IA

'VALUES FOR DCS I Control Structure Depreciation, years 1-15 (Slacre)
DATA 0, 0' Op 4' 4' ‘p 4' 4p 4

FOR IA I 1 TO NWTMS
READ DCS(IA)
NEXT IA

'VALUES FOR OCCS I Operating Costs for Control Structure (S/acre)
DATA 0, 0, 0, 0.04, 0.04, 0.04, 0.04, 0.04, 0.04

FOR IA - 1 TO NNTMS
READ OCCS(IA)
NEXT IA

137

'VALUES FOR OCP I Operating Costs Associated with Pump
DATA 0, 0, 0, 0.83, 0.83, 0.83, 0.98, 0.98, 0.98

FOR IA - 1 To NWTNS
READ OCP(IA)
NEXT IA

'VALUES FOR PLABOR I Labor Cost for Irrigation (S/acre)
DATA 0, 0, 0, 3.4, 3.4, 3.4, 3.4, 3.4, 3.4

FOR IA . 1 TO NNTXs
READ PLABOR(IA)
NEXT IA

'VALUES FOR RPUHP I Replacement Cost for
DATA 0, 0, 0, 30, 30, 30, 50, 50, 50

Pump Falls in Year 16 (S/acre)

FOR IA - 1 To NWTNs
READ RPUMP(IA)
NEXT IA

'VALUES FOR THE YR MATRIX

DATA
DATA
DATA
DATA

1958,
1967,
1976,
1985,

1959,
1968,
1977,
1986,

1960,
1969,
1978,
1987,

1961,
1970,
1979,
1988,

1962,
1971,
1980,
1989,

1963,
1972,
1981,
1990

1964,
1973,
1982,

1965, 1966
1974, 1975
1983, 1984

FOR I . 1 TO NYRS
READ XR(I)
NEXT I

'VALUES FOR THE D2 MATRIX

DATA
DATA
DATA
DATA
DATA
DATA
DATA 59.8,
DATA
DATA
DATA
DATA
DATA
DATA
DATA

100.0
99.6
100.0, 100.0
99.6, 100.0
99.7, 100.0
100.0, 65.0

100.0, 75.9,
52.8, 100.0,
100.0, 100.0,
100.0, 73.8,
60.1, 100.0,
86.8,
100.0,

84.0, 96.8,
69.7, 93.6,
100.0,
100.0,
100.0,
100.0,
100.0

0
‘
000000

FOR JE - 1 TO 2
FOR IE . 1 TO NYRS
READ D2(IE, JE)
NEXT IE

NEXT JE

'VALUES FOR THE D3 MATRIX
DATA 100., 76.7, 84.5, 97.0,
DATA 52.9, 100.0, 70.1, 94.0, 99.6
DATA 100., 100.0, 100., 100., 100.
DATA 100.0, 74.0, 100.0, 99.6, 100.0
DATA 60.5, 100.0, 100.0, 99.8, 100.0
DATA 98.8, 87.1, 100.0, 100.0, 65.2
DATA 59.9, 100.0, 100.0

100.0

DATA
DATA
DATA
DATA
DATA
DATA
DATA

FOR JE I

NEXT

138

0.
0.
0.
0.
0.
0.

0.,
0.,
00'
0.,
00'
00'
00'

0.,
0.,
0.,
0.,
00'
00'
0.,

0.,
0.,
0.,
0.,
0e,
00’
0.

00'
00'
00'
0.,
00'
00'

1 TO 2

FOR IE . 1 To NYRS
READ D3(IE, JE)
NEXT IE

33 .

'VALUES FOR THE D6 MATRIX

DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA

FOR JE I 1

NEXT

100.0,
55.6,

100.0,
100.0,

82.2,
100.0,

95.9,

76.1,
100.0,
90.2,
100.0,

86.3, 97.9, 100.0
72.4, 95.3, 99.6
100.0, 100.0, 100.0
100.0, 98.0, 100.0
100.0, 100.0, 100.0

100.0, 100.0, 67.2

100.0

C
‘
OOOOOO

TO 2
FOR IE I 1 TO NYRS
READ D6(IE, JE)
NEXT IE

JE

'VALUES FOR THE 82 MATRIX

DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA
DATA

100.0, 93.8, 100.0,
100.0, 99.1, 100.0, 100.0, 98.0
82.0, 92.2, 95.0, 99.9, 91.5
100.0, 99.0, 97.5, 85.6, 99.9
100.0, 91.5, 92.7, 100.0, 100.0

98.0, 100.0, 100.0, 100.0, 100.0

100.0, 99.9

100.0,
16.92,
26.31,
17.06,
20.46,
24.06,
20.69,
27.69,

100.0,
20.78,
19.06,
14.32,
19.33,
19.89,
22.00,
16.25,

FOR JE I 1 TO 2

NEXT

FOR IE I 1 TO NYRS

100.0

20.71,
24.36,
16.91,
19.27,
15.57,
19.79,
18.18

READ 82(IE, JE)
NEXT IE

JE

19.15,
24.95,
19.92,
20.57,
18.40,
17.24,

18.37
19.25
15.12
18.97
18.67
27.73

'VALUES FOR THE S3

DATA 100.0, 96.0,
DATA 100.0, 100.0,
DATA 91.9, 94.3, 9
DATA 100.0, 96.7,
DATA 100.0, 99.5,
DATA 99.6, 100.0,
DATA 100.0, 99.7,
DATA 14.40, 17.70,
DATA 23.81, 15.46,
DATA 13.51, 10.50,
DATA 17.76, 17.74,
DATA 22.08, 16.60,
DATA 17.09, 18.80,
DATA 25.41, 12.07,

FOR JE'I 1 TO 2
FOR IE I 1

139

MATRIX

100.0, 100.0, 99.7
100.0, 100.0, 98.8
5.0, 100.0, 93.1
99.9, 98.2, 100.0
99.7, 100.0, 100.0
100.0, 100.0, 100.0
100.0
19.24, 16.60, 14.97
22.42, 22.36, 16.14
13.78, 16.45, 12.27
15.70, 17.65, 15.42
12.39, 14.95, 15.94
16.93, 13.83, 25.53
15.66

TO NYRS

READ 83(IE, JE)

NEXT IE
NEXT JE

'VALUES FOR THE 86

DATA 100.0, 100.0,
DATA 86.1, 100.0,
DATA 97.5, 95.9, 1
DATA 100.0, 90.2,
DATA 94.8, 100.0,
DATA 100.0, 100.0,
DATA 64.9, 100.0,
DATA 7.83, 9.31, 1
DATA 9.89, 8.06, 1
DATA 6.17, 4.24, 7
DATA 9.38, 9.97, 7
DATA 10.61, 9.38,
DATA 9.13, 10.02,
DATA 3.80, 5.46, 9

FOR JE I 1 TO 2
FOR IE I 1

MATRIX

100.0, 100.0, 99.7
96.9, 96.8, 99.7
00.0, 100.0, 99.8
100.0, 98.0, 100.0
100.0, 100.0, 100.0

100.0, 100.0, 68.6
100.0
1.10, 10.17, 7.27
1.03, 10.36, 8.54
.37, 10.40, 6.86
.35, 8.85, 9.38
7.00, 9.02, 9.33
10.20, 7.44, 7.38
.14

TO NYRS

READ S6(IE, JE)

NEXT IE
NEXT JE

PRINT #1,
PRINT #1,
PRINT #2,
PRINT #3,

EASE SCENARIO"
MONTE CARLO SIMULATION TO GENERATE ENPV,
HISTOGRAM GENERATION FOR ALTERNATIVE WTMS
MONTE CARLO SIMULATION TO GENERATE CDFS OF WTMS INVESTMENTS"

SDNPV OF OPTIONS”
INVESTMENTS”

'OUTERMOST LOOP - CONDITIONS FOR EACH WTMS ARE INITIALIZED

FOR IA I 1 TO NNPV * NRUNS

R . RND
N - (NYRS - 1)
X(IA) . FIX(N)

NEXT IA

'Creates an Array of 3000 Random

'Numbers from 1-33 which are used
* R 'in referencing a yield, volume
+ 1 'pair for the Monte Carlo runs.

140

FOR IA I 1 TO NWTMS
FOR JE I 1 TO 2

NEXT JE

PNAT . PWAT(IA)
vrc - VPC(IA)
POTY - POTY(IA)

FOR IE - 1 TO NYRS
IF IA - 1 THEN A(IE,
IF IA - 2 THEN A(IE,
IF IA - 3 THEN A(IE,
IF IA - 4 OR IA . 7
IF IA . 5 OR IA . 8
IF IA - 6 OR IA . 9
NEXT IE

ONTNs - ONTNs(IA)

occs . OCCS(IA)
OCP - 0CP(IA)
TOT s o

TDIFF . 0

FOR I I 1 TO 7

COUNT(I) . 0

NEXT I

FOR I I 1 TO NWTMS

ENPV(I)
NEXT I

I 0

FOR I . 1 TO NWTMS
SDNPV(I) . 0

NEXT I

PRINT #1,
PRINT ’2,

"WTMS”; IA
“WTMS”;

IA

'MONTE CARLO LOOP

FOR I1 I 1 TO NRUNS

'RESET XNPV,

XNPV I -CWTMS
XNR I 0

DNR I 0

SUM1 I 0

SUM2 I 0

DP I 0

XNR,

DNR,

'For each WTMS, the correct yield,
'volume matrix is chosen for the
'NPV calculations.

I D2(IE, JD)

I D3(IE, JD)

I D6(IE, JE)

A(IB, JE) I s2(IE,
A(IE, JE) I S3(IE,
A(IE, JE) . S6(IE,

'Selects correct prices and costs

'for alternate WTMSs.

'Sums NPVs for ENPV calculations.
'Difference between NPV and ENPV for
'SD calculations.

'Resets count to zero for histogram
'percentages.

'Resets ENPV to zero before
'each WTMS.

'Resets SDNPV to zero before
'each WTMS.

SUMl AND DF TO ZERO BEFORE EACH SIMULATION

'initial cost WTMS (year zero)
'undiscounted net revenue
'discounted net revenue

'first sum in NPV calculation
'second sum in NPV calculation
'discount factor

'INNERMOST LOOP FOR CALCULATING NPV OVER A 30-YEAR INVESTMENT HORIZON

FOR M I 1 TO NNPV

IF M (I 7 THEN

DPUMP I DPUMP(IA)

'Establishes correct depreciation
'periods for pump, well, tile, CS.

ELSEIF M > 7 AND M <I 15 THEN

DPUMP I 0

ELSEIF M > 15 AND M <I 22 THEN

DPUMP I DPUMP(IA)

141

ELSE

DPUMP I 0
END IF
IF M <I 15 THEN DCS I DCS(IA) ELSE DCS I 0
IF M (I 15 THEN DWELL I DWELL(IA) ELSE DWELL I 0
IF M (I 15 THEN DTILE I DTILE(IA) ELSE DTILE I 0
IF M I 16 THEN RPUMP I RPUMP(IA) ELSE RPUMP I 0

PLACE‘ I ((11 * NNPV) - NNPV) + M 'References an element in the yield
RYLD I A(R(PLACE‘), 1) / 100 'volume matrix.

VOL . A(R(PLACE%), 2)

YEAR I YR(M)

YLD I RYLD * POTY 'conversion of rel yld to bu/acre
VPBC I YLD * PBC 'variable per bushel

'production costs
VIC I VOL*PWAT+OCCS+OCP+PLABOR 'variable cost of irrigation

SUMl I (POUT * YLD - VIC - VPC - VPBC - DWELL - DPUMP - DCS - DTILE -
OCI * (VIC + VPC + VPBC)) * (1 - TB)
SUM2 I DWELL + DPUMP + DCS + DTILE 'Depreciation cost, which are

XNR I SUMl + SUMZ - RPUMP 'subtracted out in SUMl for tax
DF I (1 + ATRR) “ M 'purposes must be added back in
DNR I XNR / DF 'to reflect actual cash flow.

XNPV I XNPV + DNR
NEXT M

'END OF NPV LOOP

XNPVRUN(Il) I XNPV 'Array of NPV for prob dist.
TOT I TOT + XNPV 'For ENPV calculations.
NEXT I1

'END OF MONTE CARLO LOOP
ENPV(IA) - TOT / NRUNs 'Array of ENPV for all WTMSS

FOR I I 1 TO NRUNS 'Calculation of SD
DIFF I (XNPVRUN(I) - ENPV(IA)) ‘ 2
TDIFF I TDIFF + DIFF

NEXT I

SDNPV(IA) I (TDIFF / NRUNS) * .5 'Array of SD for all WTMSs

PRINT #1, '”
PRINT #1, USING ' ENPV I SSIfffi”; ENPV(IA);
PRINT #1, USING ” SDNPV ' $$i#f#”; SDNPV(IA)
PRINT #1, ""

'CALCULATION OF HISTOGRAM VALUES
'DETERMINING MAX AND MIN VALUES OF NPV (SPREAD)

XXINVAL . XNPVRUN(1)
XMAXVAL . XNPVRUN(1)

FOR I - 1 To NRUNs

IF XNPVRUN(I) < XMINVAL THEN XMINVAL . XNPVRUN(I)
IF XNPVRUN(I) > XMAXVAL THEN XMAXVAL - XNPVRUN(I)
NEXT I

142

PRINT ’2, USING “MINIMUM VALUE I ffff’”: XMINVAL
PRINT #2, USING ”MAXIMUM VALUE I I####"; XMAXVAL

DIM HIST(7)
BARWIDTH I (XMAXVAL - XMINVAL) / 7

PRINT ’2, "
PRINT #2, USING ”BARWIDTH: ##F”; BARNIDTH

FOR I - 1 TO 7
EIST(I) - XMINVAL + BARWIDTH * (I - 1)
NEXT I

PRINT #2, "LOWER LIMITS ON HISTOGRAM EARS”
FOR I I 1 TO 7

PRINT #2, ”BAR"; I;

PRINT #2, USING ”LOWER LIMIT: ##f’"; HIST(I)
NEXT I

FOR J I 1 TO NRUNS

I I 8

100 I I I - 1

IF XNPVRUN(J) >I HIST(I) THEN COUNT(I) I COUNT(I) + 1 ELSE 100
NEXT J

PRINT #2, "HISTOGRAM: PERCENTAcEs IN EACH BARz"

FOR I a 1 TO 7

PRINT #2, ”PERCENTAGE IN BAR”; I; (COUNT(I) / NRUNS)
NEXT I

'SORT TREATMENTS BY NPV IN DESENDING ORDER

DO
SWAPS% - FALSE‘
FOR I . 1 TO (NRUNS - 1)
IF XNPVRUN(I) < XNPVRUN(I + 1) THEN
SWAP XNPVRUN(I), XNPVRUN(I + 1)
. SWAPS% . I
END IF
NEXT I
LOOP WHILE SWAPS!

PRINT #3, ”Full sort results are for WTMS"; IA; ": ”
FOR I I 1 TO NRUNS

PRINT #3, USING "#####"; XNPVRUN(I)

NEXT I

NEXT IA

'END OF OUTERMOST LOOP
END

 

APPENDIX E

DRAINMOD DATA INPUTS

*t***********t*******t***t**t****************************i**************
d r a I n m o d

version: north carolina micro 4.05
Last update: sept 1991
language: ms fortran v 5.0

Copyright (c) 1990, north carolina state university
all rights reserved

Drainmod is a field-scale hydrologic model developed for
the design of subsurface drainage systems. The model was
developed by researchers at the dept. of biological and
agricultural engineering, north carolina state university
under the direction of R. W. Skaggs.

************************************************************************
*******************

* D r a I n m O d *
*******************

data read from input file: c:\dm40\input40\dr2k12.Lis

title of run

************

dr, 20 ft tile spacing, kilmanagh soil I kilmancm, flint weather,
dry slope 1.05, Net slope I .68, Sf I 1.25, Plant days I 8, 1958-90

climate inputs
******* ******

description (variable) value

 

file for raindata .............. C:\dm40\weather\fnt5891.Rai
file for temperature/pet data .. C:\dm40\weather\fnt5890.Tem

rainfall station number..........................(Rainid) 202846
temperature/pet station number...................(Tempid) 202846
starting year of simulation..................(Start year) 1958
starting month of simulation................(Start month) 1
ending year of simulation......................(End year) 1990
ending month of simulation....................(End month) 12
temperature station latitude...................(Temp lat) 43.03
Heat index..........................................(Hid) 40.00

ET multiplication factor for each month
1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

143

144

DRAINAGE SYSTEM DESIGN

**********************

*** CONVENTIONAL DRAINAGE ***

JOB TITLE: DR, 20 ft tile spacing, Kilmanagh Soil I RilmanCM, Flint
Weather, NET SLOPE I .68, SF I 1.25, PLANT DAYS I 8, 1958-90

 

STMAX I 2.50 CM SOIL SURFACE
__/) /)__
: :
z :
ADEPTH I152. CM DDRAIN I 102 CM

 

o — SDRAIN = 610 CM ----------- o -

\
EFFRAD = .51 CM
HDRAIN = 30. CM

IMPERMEABLE LAYER

 

////////////////[//////////////////////////////////////////////////

DEPTH SATURATED HYDRAULIC CONDUCTIVITY
(CM) (CM/HR)
.0 - 74.0 30400
112.0 - 152.0 .150

depth to drain I 102.0 Cm

effective depth from drain to impermeable layer I 29.8 Cm

distance between drains I 610.0 Cm

maximum depth of surface ponding I 2.50 Cm

effective depth to impermeable layer I 131.8 Cm

drainage coefficient(as limited by subsurface outlet) I .95 Cm/day
actual depth from surface to impermeable layer I 152.0 Cm

surface storage that must be filled before water can move to drain I
.50 Cm

factor -g- in kirkham eg. 2-17 I13.01

width of ditch bottom I 1.0 Cm

side slope of ditch (horiz:vert) I .10 : 1.00

Initial water table depth I 38.0 Cm

DEPTH OF WEIR FROM THE SURFACE

 

DATE 1/ 1 2/ 1 3/ 1 4/ 1 5/ 15 6/ 15
WEIR DEPTH 102.0 102.0 102.0 102.0 102.0 102.0

DATE 7/ 1 8/ 15 9/ 1 10/ 1 11/ 1 12/ 1
WEIR DEPTH 102.0 102.0 102.0 102.0 102.0 102.0

145

SOIL INPUTS

***********

VOID VOLUME WATER TABLE DEPTH

(CM) (CM)
.0 .0
1.0 27.6
2.0 39.3
3.0 49.1
4.0 59.0
5.0 69.5
6.0 80.2
7.0 91.1
8.0 101.4
9.0 111.8
10.0 121.9
11.0 131.1
12.0 140.4
13.0 149.7
14.0 158.2
15.0 166.7
16.0 175.1
17.0 183.6
18.0 192.1
19.0 200.4
20.0 206.9
21.0 213.4
22.0 219.8
23.0 226.3
24.0 232.8
25.0 239.2
26.0 245.7
27.0 252.2
28.0 258.7
29.0 265.1
30.0 271.6
35.0 303.9
40.0 336.3
45.0 368.6
50.0 400.9
60. 465.6
70.0 567.5
80.0 711.7

90.0 855.8

146

SOIL WATER CHARACTERISTIC VS VOID VOLUME VS UPFLUX

HEAD WATER CONTENT VOID VOLUME UPFLUX
(CH) (CH/CM) (CH) (CH/HR)
. .4760 .00 .5000
10.0 .4520 .13 .5000
20.0 .4280 .52 .1599
30.0 .4040 1.17 .0527
40.0 .3800 2.07 .0237
50.0 .3780 3.09 .0131
60.0 .3760 4.10 .0048
70.0 .3742 5.05 .0029
80.0 .3725 5.98 .0017
90.0 .3707 6.90 .0011
100.0 .3690 7.86 .0009
110.0 .3670 8.83 .0007
120.0 .3650 9.80 .0005
130.0 .3629 10.88 .0004
140.0 .3609 11.96 .0004
150.0 .3589 13.03 .0003
160.0 .3569 14.21 .0002
170.0 .3549 15.39 .0002
180.0 .3529 16.57 .0001
190.0 .3508 17.75 .0001
200.0 .3488 18.93 .0000
210.0 .3468 20.48 .0000
220.0 .3448 22.02 .0000
230.0 .3428 23.57 .0000
240.0 .3408 25.12 .0000
250.0 .3387 26.66 .0000
260.0 .3367 28.21 .0000
270.0 .3347 29.75 .0000
280.0 .3327 31.30 .0000
290.0 .3307 32.85 .0000
300.0 .3287 34.39 .0000
350.0 .3213 42.12 .0000
400.0 .3191 49.85 .0000
450.0 .3169 57.58 .0000
500.0 .3147 65.31 .0000
600.0 .3104 72.25 .0000
700.0 .3060 79.19 .0000
800.0 .3017 86.13 .0000
900.0 .2973 93.06 .0000
GREEN AMPT INFILTRATION PARAMETERS
W.T.D. A 3
(CM) (CH) (CH)

.000 .000 3.300

10.000 .440 3.300

20.000 .890 3.300

40.000 1.710 3.300

60.000 1.770 3.270

80.000 1.840 3.270

100.000 1.890 3.270

150.000 4.050 3.270

200.000 4.050 3.270

1000.000 4.050 3.270

147

second

period
3.40
1.30
2.00

9/ 1

11/ 1
a
18

TRAFFICABILITY
**************
first
requirements period
Iminimum air volume in soil (cm): 3.40
-Maximum allowable daily rainfall(cm): 1.30
-Minimum time after rain to continue tilling: 2.00
working times
-date to begin counting work days: 4/20
-date to stop counting work days: 6/ 1
-first work hour of the day: 8
-last work hour of the day: 20
crop
*ttt
soil moisture at crop wilting point I .22
High water stress: begin stress period on S/ 1
end stress period on 9/ 1
crop is in stress when water table is above
drought stress: begin stress period on 5/ 1
end stress period on 9/ 1
MO DAY ROOTING DEPTH(CM)
1 1 3.0
5 7 3.0
S 25 5.0
6 8 20.0
6 22 35.0
7 13 40.0
8 9 45.0
9 10 30.0
10 15 10.0
10 20 3.0

30 cm

148

YIELD INPUTS
*ttttttttttt

last planting day without yield loss (JLAST):
length of growing season (IGROW)
1st planting day reduction factor (PDRF)

days using lst planting delay fact (DELAYl)

2nd planting day reduction factor (PDRF2)

total days of work before planting (REQWRK)

.0000
.5000
1.7500
1.3000
.5000

10W: 30
IOH: 11
SI : 11.160000
D : -1.170000
E : 5.800000E-02
F0 : I5.0000003-04
YI : 100.000000
SF : 1.250000
YRMAX : 0.000000E+00
YSLOPE: 1.050000
YRDMAX: 100.000000
DSLOPE: 6.8000003-01
PD : 121
IGR: 105
SDF: 1
IPS(I),IPE(I),CSD(I),II1,IOH
0 9 .2000
10 19 .2000
20 35 .2200
36 49 .2800
50 59 .3200
60 69 .2800
70 79 .1900
80 89 .1200
90 99 .0800
100 104 .0400
105 105 .0200
CSI(I),II1,IOW
.0000 .0000
.0000 .5000
1.0000 1.0000
1.3000 1.3000
1.2000 1.0000
.0000 .0000

**> Total

simulation timeI

.0000

4.333 minutes.

130

105

6.000000E-01

.0000

1.0000
2.0000
1.3000

.0000
.0000

22.000000
1.800000
8.000000

.0000

1.0000
2.0000
1.3000

.0000
.0000

149

 

dr2k12.lil

 

c:\DN40\INPUT4O\DR2X12.OEN
c:\DH40\INPUT40\RILCHX12.sIN
C:\DM40\INPUT40\CN105K12.YIN

dr2k12.gen

 

*** Job Title ***
DR, 20 ft tile spacing, Kilmanagh Soil I RilmanCM, Flint Weather, dry
slope 1.0
WET SLOPE I .68, SF I 1.25, PLANT DAYS I 8, 1958-90
*** Printout and Input Control ***
3 l 0 C:\DM40\OUTPUT40\
*tt Climate tit
202846, C:\DM40\WEATHER\FNT5891.RAI
202846 C:\DM40\WEATHER\FNT5890.TEN
1958 1 1990 12 4303 40
1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
*** Drainage System Design ***
1

102.00 29.84 610.00 2.50 0.95 0.50 13.01
38.00
0 , 0 , 0
0 , 0 , 0 , 0
0 , 0 , 0 , 0 ,
1.00 0.10
1107 1107 1107 110715 7615 51 1 5115 76 1107 1107 1107 1107
iii $0113 ***
152.00 0.51
74. 3.40 112. 2.79 152. 0.15
99
*** Trafficability ***
420 6 1 820 3.
9 111 1 818 3.
iii Crop *t*
0.220
5 1 9 1 30.00
5 1 9 1
11
1 l 3.0 5 7 3.0 525 5.0 6 8 20.0 622 35.0 713 40.0 8 9 45.0
910 30.0
1015 10.0 1020 3.0 1231 3.0
*** Wastewater Irrigation ***
0 0 0 365 0 0
0 0 0 0 0 0 0 0
0.00000 0.00000 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
0.00 0.00

0

F‘H
wt»
Raw
00

4
4

cn105k12.yin

130 105 0.6000 8.0000 1.8000 22.0000
3011 11.1600 -1.1700 0.0580 -0.0005 100.0000 1.2500

100.000 1.050 100.000 0.680 121 105 1

0 90.20 10 190.20 20 350.22 36 490.28 50 590.32 60 690.28 70 790.19
80 890.12

90 990.081001040.041051050.02
0.000.000.000.000.000.000.500.501.001.001.001.001.752.002.001.301.301.30
1.301.30
1.201.000.500.000.000.000.000.000.000.00

kilcmk12 . sin

4 LAYER- XILMANAGH, HURON CO. MICHIGAN

1120

0.47600
0.38000
0.37600
0.36900
0.32200
0.29300
0.26600
0.25700
0.24800
0.24100
0.23700

0.0000

3.0000

6.0000

9.0000
12.0000
15.0000
20.0000
25.0000
30.0000
35.0000
40.0000
45.0000
60.0000
75.0000
90.0000

120.0000
150.0000
200.0000
500.0000

1000.0000

10

0.00
10.00
20.00
40.00
60.00
80.00

100.00
150.00
200.00
1000.00

0.0
-40.0
-60.0

-100.0
-333.0
-1000.0
-2000.0
-3000.0
-5000.0

-10000.0
-15000.0

0.0000
0.0120
0.0470
0.1050
0.1870
0.2930
0.5200
0.8130
1.1700
1.5910
2.0720
2.5820
4.1020
5.5200
6.8960
9.7970
13.0340
18.9320
65.3140

100.0000

0.00
0.44
0.89
1.71
1.77
1.84
1.89
4.05
4.05
4.05

0.5000
0.5000
0.5000
0.5000
0.5000
0.3417
0.1599
0.0869
0.0527
0.0340
0.0237
0.0172
0.0048
0.0020
0.0011
0.0005
0.0003
0.0000
0.0000
0.0000

3.30
3.30
3.30
3.30
3.27
3.27
3.27
3.27
3.27
3.27

150

APPENDIX C

DRAINAGE-ONLY SITE DESIGN
by Dr. Harold Belcher
Department of Agricultural Engineering
Michigan State University

Subsurface Drain
7° 0"“ Pipe Lateral

  
   

O

.5

_l

b

3

E ‘5

5 0
8:
{a
a:
O
3.9
030.

151

APPENDIX D

SUBIRRIGATION SITE DESIGN
by Dr. Harold Belcher
Department of Agricultural Engineering
Michigan State University

Subsurface Drain Subsurface Draln
Pipe Main Plpe Lateral

To Outlet

Head Control Stand

 

 

152

APPENDIX E

AVERAGE MONTHLY TEMPERATURE AND PRECIPITATION FOR BAD AXE, HARBOR BEACH,
AND FLINT, MICHIGAN FOR THE PERIOD 1951-1980

 

 

 

 

 

 

 

 

 

 

 

153

 

 

 

I Station Temperature (°F) Precipitation I
(in)
Bad Axe Avg Daily Avg Daily Average Average
I Max Min
January 29.1 14.3 21.7 1.86
February 31.2 14.3 22.7 1.87
March 39.5 22.2 30.9 2.30
April 55.0 33.8 44.4 2.66
May 67.0 43.0 55.0 2.60
June 77.4 53.5 65.4 2.86
July 81.5 57.5 69.5 3.01
August 80.2 56.4 68.3 2.66
September 72.3 49.9 61.1 2.48
October 62.0 41.3 51.7 2.39
November 45.8 31.0 38.4 2.39
December 33.4 20.3 26.9 2.09
Yearly Avg 56.2 36.5 46.3 29.17
Harbor Avg Daily Avg Daily Average Average fl
Beach Max Min
January 28.7 15.3 22.0 2.66
February 30.5 15.8 23.2 2.31
March 37.7 23.2 30.5 2.47
April 52.0 34.1 43.0 2.84
May 63.0 42.7 52.8 2.63
June 73.9 53.2 63.5 3.18
July 78.2 58.7 68.5 3.22
August 77.3 58.4 67.9 3.16
September 70.3 51.9 61.1 2.75
October 60.2 42.8 51.5 2.66
November 45.4 32.2 38.8 2.89
December 33.3 21.3 27.3 3.17
Yearly Av 54.2 37.5 45.8 33.84
Flint Avg Daily Avg Daily Average Average
Max Min
January 31.9 17.1 24.5 1.63
February 32.8 16.7 24.8 1.76
March 41.3 24.1 32.7 2.20
April 55.9 34.7 45.3 2.85
May 68.0 44.7 56.4 3.16
June 78.5 54.7 66.6 3.32
July 83.5 58.9 71.2 2.86
August 81.6 57.6 69.6 3.43
September 73.4 50.5 62.0 2.53
October 62.1 40.6 51.4 2.09
November 46.3 30.1 38.2 2.05
December 34.6 20.6 27.2 1.70
57.5 37.5 47.5 29.58

APPENDIX F

DRAINMOD WATER BALANCE VERIFICATION
Dr. Harold Belcher
Department of Agricultural Engineering
Michigan State University

Introduction:

DRAINMOD simulation results have been compared to observed data at a number of
locations including North Carolina, South Carolina, Louisiana, Florida, Georgia, Iowa,
and Ohio [notez references are probably in DRAINMOD USER MANUAL]. In all
cases, it was reported the simulated water balance results (water table depth and/or
subdrain discharge) were reasonably close to observed data. To evaluate applicability of
the model to Michigan climate and poorly drained soils, observed water table depth data
from a water table management research site near Bannister Michigan was compared to
Drainmod simulated water table depths for two years of record.

Bannister Site:

The Bannister site is described by Belcher, 1990. Soil at the site is classified as a
Ziegenfuss silty clay loam and has particle size gradation and hydraulic properties similar
to typical Saginaw Bay area shallow water table, poorly drained soils (see Table 1).

Soil property inputs for DRAINMOD resulted from application of the DRAIN MOD
”soilprep" computer model using the field measured soil water characteristic data as
follows:

Lateral saturated hydraulic conductivity values for each soil layer used for DRAINMOD
(see Table 1) are the mean of the field determined values reported by Fogiel and
Belcher (1990) for the areas that do not include sand. The values used are the average
of velocity permeameter (Merva, 1987) lateral measurements, 36 each at depths 0.45 m,
0.60 m and 0.75 m. For the ”soilprep" model, the lateral conductivities were reduced by
50% to approximate vertical saturated hydraulic conductivity for each layer.

Results:

The results of this study are presented in Figures 1, 2 and 3 for the 1986 growing season
and subdrain lateral spacings of 6 m, 12 m and 18 m, respectively and Figures 4, and 5
for the 1987 growing season and 6 m and 18 m subdrain lateral spacings. The reader is
referred to Belcher, 1990 for a detailed description of the Bannister site, instrumentation
and observed water table figures.

154

H5

Table 2. Properties of a Ziegenfuss soil. typical at the Bannister, Michigan
research site (Rosek, 1992).

 

Depth Sand Clay Organi
c
Matter

inches

 

 

 

 

 

3-1 25-38 28 36 36 0.7 1.50 1.1
892

B-l CI 38-52 20 36 44 0.7 1.56 1.1
lB-l c2 52-60 0 37 63 0.7 1.71 0.1*

 

 

 

 

 

 

 

 

 

* Values assumed for DRAINMOD simulations.

 

 

 

 

 

750 no

A

e l

E 5am; >501)

V

g 25.0- -25.o
0.0; - - LIL- -‘fi IV 3.0
:65 40,:

lo I I manusemusoum

’E‘ -— simulated

V 30.31 103

§ 1

'3

g «4

.1

U

' 3

.-

.-

<

I
”"11 -fij—f ' r'fif' rfififiv ' I"' V ''''' ”-5

m :5: m m 200 m 223 2.): m

 

 

DAYS FROM STAR? OF 1986
MISTER SITE OBSERVATION WELL WC2L1

 

 

 

Figure 1. Observed and predicted water table elevation (m)
for subdrains spaced at 2 m for 1986 growing season at the
Bannister site.

156

Table 3. Volumetric water contents (cm/cm) at various soil tensions (cm) for

typical Ziegenfuss soil at Bannister, Michigan research site (Rosek, 1992).

    
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 cm
0.5241 0.4646 0.4604 0.4567 0.442 0.4304 7‘
0.5011 0.4343 0.4299 0.4258 0.3981 0.388
0.5151 0.4486 0.4451 0.4362 0.388 0.3816
0.5212 0.4541 0.4492 0.4318 0.3906 0.3786
B-l C2 0.5647 0.5013 0.4972 0.4906 0.4483 0.4334 J
Sample 2000 cm 3000 cm 5000 cm 10000 15000
C111 CID
B-1 Ap 0.415 0.4074 0.3864 0.3805 0.368
B-1 Bgl 0.3769 0.3698 0.3649 0.3589 0.3475
B-1 892 0.3686 0.3606 0.3469 0.3412 0.3305
B-l C1 0.3608 0.3527 0.3415 0.3361 0.3269
B-1 C2 0.4154 0.4052 0.3855 0.3802 0.3688
[_l—=L_____ L l.

 

 

 

 

MW

 

RAIN (mm)

214‘

wuss mu: summon (m)

 

  
   

25.0‘ I I I - 25.0
0.0“ -0.0

20‘ m ”.07!!!“ 30.10 n

:q

Fwfl

> 20.8

 

 

 

1%

T

18: m 206 2i: 228 236 2
DAYS FROM sum 0F1698
BANNISTER SITE OBSERVATION WELL W84M2

 

 

Figure 2. Observed and predicted water table elevation (m)

for subdrains spaced at 12 m for

Bannister site.

1986 growing season at the

 

RAIN (mm)

WAIER TABLE ELEVATION (m)

 

A .Ll- I I

V V Y

200‘ ELEV: 30.18mte 30.48:!)

 

 

 

 

' v
178

1‘81 Wis :66 I V Tie V V25: 2511' V
DAYS FROM START OF 1956
BANNISIER SITE OBSERVAIION WELL WCGMI

 

Figure 3. Observed and predicted water table elevation (m)
for subdrains spaced at 18 m for 1986 growing season at the

Bannister site.

 

 

158

 

 

 

 

 

73.0- 750
A I
E 1
5 50m -so.o
V 1
3 no; I .m
0:: V - V v V I L v- W'OD
30.0 '30.0
A
E
v
z
8
3
—J
U
‘3
3
K
t‘.’
(
3
29.3 r - v v - v v v v w 210
III? 192 202 212 222 232 2‘2

 

DAYS FROM START OF 1987
BANNISTER STTE OBSERVATION WELL WDZLI

 

 

 

rigure 4. Observed and predicted water table elevation (m)
for subdrains spaced at 6 m for 1987 growing season at the
Bannister site.

 

75.0 - 75 O

 

 

 

 

 

?

E 50.0 woo

v

g 25.0 I >210
0,3r I - Y v I. Y‘ WLOO
”.0‘ Jon

0 oh” I zonamsoumu ”Jam

A {

E

V

E

U

U

i

'—

K

5

3
”-3 v v v r v v 1 v r v V V v 29.0

Ill? ‘92 202 2‘2 222 232 2‘?

DAYS FROM START OF 1987
BANNISTER SITE OBSERVATION WELL WDGLI

 

 

 

Figure 5. Observed and predicted water table elevation (m)
for subdrains spaced at 16 m for 1987 growing season at the
Bannister site.

159

Discussion:

The predicted water table elevations in Figures 1 through 5 result from DRAINMOD
version 4.0 without modification of inputs to calibrate the model.

DRAINMOD allows a single water table control weir setting per month. During both
the 1986 and 1987 seasons, the Bannister water table control weirs were lowered
following selected rainfall events thus weir settings were sometimes altered more often
than one time in a month. When this occurred, a variance between observed and
predicted water table depth is to be expected. This is best illustrated by looking at
Figure 2 at 188 days.

The water table at the Bannister site was observed by monitoring instrumentation that
allowed hourly observation. During peak evapotranspiration days, it was observed that
the water table varied as much as 15 cm from morning to mid-afternoon. DRAINMOD
provides a daily water table depth output which does not provide the observed hourly
fluctuation.

The water table depth instrumentation at Bannister has a limited operating range.
Examination of the figures indicate that the actual water table depth sometimes
exceeded the upper limits. Thus, DRAINMOD data showing a higher water table
elevation than was observed are not unexpected.

Considering the preceding discussion, the results as provided by Figures 1 through 4
indicate strongly that DRAINMOD does accurately model the change in water table
depth with time for a poorly drained soil in Michigan.

References:
Belcher, H. W. 1990. Water table management to maximize the economic efficiency of
biomass production. PhD. Dissertation. Agricultural Engineering Department, Michigan

State University.

Fogiel, A. C. and H. W. Belcher. 1990. Watertable management effects on soil structure.
ASAE Paper No. 90-2604. ASAE, St. Joseph, MI.

Rossek, M. 1992. Personal Communication, Michigan State University, E. Lansing, MI.
Merva, G. E. 1987. The velocity penneameter technique for rapid determination of

hydraulic conductivity in-situ. Proceedings of the Third International Workshop on Land
Drainage. Ohio State University, Columbus, OH.

APPENDIX 0

INVESTMENT COSTS

Contractor Estimates For Investment Options

DR20
Cost Estimates for Modifying a Conventional Drainage System
from 60-ft Tile Spacings to 20-ft Spacings

Diameter Quantity
Item of Work (inches)(# or ft) Item Cost Item Cost Item Cost Avg Cost

 

Laterall 4 58660 $20,444 $22,984 $19,758 $21,062

 

$20,444 $22,984 $19,758 $21,062

DR30
Cost Estimates for Modifying a Conventional Drainage System
from 60-ft Tile Spacings to 30-ft Spacings

Diameter Quantity
Item of Work (inches)(# or ft) Item Cost Item Cost Item Cost Avg Cost

 

Laterals 4 29330 $10,472 $11,692 $10,079 $10,748

 

$10,472 $11,692 $10,079 $10,748

DR60
Existing System: No Costs

81208 or SIZOW

Cost Estimates for Modifying a Conventional Drainage System
from 60-ft Tile Spacings to 20-ft Spacings and Retrofitting for
Subirrigation, Surface Water Source or Well Water Source

Diameter Quantity
Item of Work (inches)(# or ft) Item Cost Item Cost Item Cost Avg Cost

 

Head Stands 3 $1,900 $2,600 $1,135 $1,878
Irrigation Inlets 3 $400 $400 $250 $350
Water Supply 6 1100 $4,120 $2,190 $3,961 $3,424
Laterals 4 58660 $19,994 $22,371 $19,392 $20,586
Mains/Submain 6 1027 $1,180 $1,199 $1,174 $1,184
(2) 8 1405 $2,158 $2,173 $2,394 $2,242
(3) 10 856 $2,190 $2,040 $2,456 $2,229
(4) 12 370 $1,197 $1,149 $1,396 $1,247

$33,139 $34,123 $32,158 $33,140

160

SI3OS or SI30W

161

Cost Estimates for Modifying a Conventional Drainage System
from 60-ft Tile Spacings to 30-ft Spacings and Retrofitting for
Subirrigation, Surface Water Source or Well Water Source

Diameter Quantity
Item of Work (inches)(# or ft)

Item Cost Item Cost Item Cost Avg Cost

 

 

Head Stands 3 $1,900 $2,600 $1,135 $1,878
Irrigation Inlets 3 $400 $400 $250 $350
Water Supply 6 1100 $4,120 $2,190 $3,961 $3,424
Laterals 4 29330 $10,022 $11,225 $9,713 $10,320
Mains/Submain 6 1027 $1,180 $1,199 $1,174 $1,184
(2) 8 1405 $2,158 $2,173 $2,394 $2,242

(3) 10 856 $2,190 $2,040 $2,456 $2,229

(4) 12 370 $1,197 $1,149 $1,396 $1,247

$23,166 $22,978 $22,479 $22,875

81608 or SI60W

Cost Estimates for Retrofitting a Conventional Drainage System at
60-ft Tile Spacings to a Subirrigation System at 60-ft Tile Spacings,

Surface Water Source or Well Water Source

Diameter Quantity
Item of Work (inches)(# or ft)

Item Cost Item Cost Item Cost Avg Cost

$1,878

 

Head Stands 3 $1,900 $2,600 $1,135
Irrigation Inlets 3 $400 $400 $250 $350
Water Supply 6 1100 $4,130 $2,350 $3,967 $3,482
Mains/Submain 6 1027 $1,190 $1,272 $1,180 $1,214
(2) 8 1405 $2,168 $2,288 $2,400 $2,285
(3) 10 856 $2,200 $2,180 $2,462 $2,281
(4) 12 370 $1,207 $1,261 $1,402 $1,290
$13,194 $12,351 $12,796 $12,781
SUMMARY OF COSTS FOR EACH INVESTMENT OPTION
DR20 DR30 DR60 SIZOS S130S $1605 SIZOH $1309 $1609
Hell 30 30 30 30 30 30 315.000 315.000 315.000
Pump 30 30 30 34.600 34.600 34.600 35.400 35.400 35.400
CS 30 30 30 32.228 $2.228 $2.228 $2.228 $2.228 $2.228

Tile 321.062 310.748

30 330.912 320.646 310.552 330.912 320.646 310.552

 

TOTAL:321.062 310.748

30 337,740 327,475 317.381 353.540 343.275 333.181

 

DEPRECIATION FOR EACN COMPONENT

DR20 DR30 DR60 S1205 $130S $1605 $1209 SI30W S1609
Hell 30 30 30 30 30 30 31.000 31.000 31.000
Pump 30 30 30 3657 3657 3657 3771 3771 3771
CS 30 30 30 3149 3149 3149 3149 3149 3149
Tile 31.404 3717 30 32.061 31.376 3703 32.061 31.376 3703
TOTAL: 31.404 3717 30 32.866 32.182 31.509 33.981 33.296 32.623

OPERATING COSTS FOR DIFFERENT COMPONENTS

162

 

 

 

 

DR20 DR30 DR60 $1205 $1305 $1605 $1209 $1309 $1609

Hell 30 3O 3O 3O 30 3O 3O 3O 30
Pump 30 30 30 333 333 333 339 339 339
CS 30 30 30 31 31 31 31 31 31
Tile 30 30 30 3O 3O 30 3O 3O 30
TOTAL 30 30 30 334 334 334 340 340 340
SUMMARY OF COSTS FOR EACH INVESTMENT OPTION 1N PER ACRE TERMS

DR20 DR30 DR60 $1208 81308 81608 SI2OW SI30W SI60W
W011 $0 $0 $0 $0 $0 $0 $375 $375 $375
Pump $0 $0 $0 $115 $115 $115 $135 $135 $135
CS $0 $0 $0 $56 $56 $56 $56 $56 $56
Tile $527 $269 $0 $773 $516 $264 $773 $516 $264
TOTAL: $527 $269 $0 $944 $687 $435 $1,339 $1,082 $830
DEPRECIATION FOR EACH COMPONENT IN PER ACRE TERMS

DR20 DR30 DR60 81205 $1308 $1608 SIZOW SI30W SIéOW
W911 $0 $0 $0 $0 $0 $0 $25 $25 $25
Pump $0 $0 $0 $16 $16 $16 $19 $19 $19
CS $0 $0 $0 $4 $4 $4 $4 $4 $4
Tile $35 $18 $0 $52 $34 $18 $52 $34 $18
TOTAL: $35 $18 $0 $72 $55 $38 $100 $82 $66
OPERATING COSTS FOR DIFFERENT COMPONENTS IN PER ACRE TERMS

DR20 DR30 DR60 81205 81308 $1608 SIZOW SIBOW SI60W
W611 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00
Pump $0.00 $0.00 $0.00 $0.82 $0.82 $0.82 $0.96 $0.96 $0.96
CS $0.00 $0.00 $0.00 $0.04 $0.04 $0.04 $0.04 $0.04 $0.04
Tile $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00
TOTAL: $0.00 $0.00 $0.00 $0.86 $0.86 $0.86 $1.00 $1.00 $1.00

APPENDIX M

SIMULATION YIELD RESULTS

H1: DR20 - Drainage Only at 20-Ft Tile Spacings

tit“!**i'kt*fitti‘tt****i***t*t***********t*t*i*******t****it*****t*i****t*****t'ki

 

---------- RUN STATISTICS ---------- time: 1/ 7/1993 9 0:25

input file: C:\DM40\INPUT40\DR2K12.L1S

parameters: free drainage and yields calculated

drain spacing . 610. cm drain depth a 102.0 cm
501 - STRESS plant plant harv. RELATIVE YIELDS (%)
excess drought date delay date excess drought delay overall

1958 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1959 .0 22.9 125 0. 230 100.0 75.9 100.0 75.9
1960 .0 15.3 125 0. 230 100.0 84.0 100.0 84.0
1961 .0 3.0 125 0. 230 100.0 96.8 100.0 96.8
1962 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1963 .0 45.0 125 0. 230 100.0 52.8 100.0 52.8
1964 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1965 .0 28.9 125 0. 230 100.0 69.7 100.0 69.7
1966 .0 6.1 125 0. 230 100.0 93.6 100.0 93.6
1967 .0 .4 125 0. 230 100.0 99.6 100.0 99.6
1968 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1969 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1970 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1971 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1972 .0 .0 125 0. 230 100 0 100 0 100.0 100.0
1973 .0 .0 125 O. 230 100.0 100.0 100.0 100.0
1974 .1 25.0 125 0. 230 100.0 73.8 100.0 73.8
1975 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1976 .0 .4 125 0. 230 100.0 99.6 100.0 99.6
1977 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1978 .0 38.0 125 0. 230 100.0 60.1 100 0 60.1
1979 .0 .0 125 0. 230 100 0 100.0 100.0 100.0
1980 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1981 .0 .3 125 0. 230 100.0 99.7 100.0 99.7
1982 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1983 .0 1.4 125 0. 230 100.0 98.6 100.0 98.6
1984 .0 12.5 125 O. 230 100.0 86.8 100.0 86.8
1985 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1986 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1987 .0 33.3 125 0. 230 100 0 65.0 100.0 65.0
1988 .0 38.3 125 0. 230 100.0 59.8 100.0 59.8
1989 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1990 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
AVG 0 8 2 125 0 230. 100.0 91.4 100.0 91 4

1Ei3

HZ: DR30 - Drainage Only at 30-Ft Tile Spacings

*Rfi‘tfiitfifiififitfttfiittfi'ttttttittit*ifitWittt*ttttttktttfiit*tttttttttttttt*****titit

---------- RUN STATISTICS ---------- time: 1/ 7/1993 9 0:29
input file: C:\DM40\INPUT40\DR3K12.LIS
parameters: free drainage and yields calculated

drain spacing - 914. cm drain depth - 102.0 cm

 

 

SDI - STRESS plant plant harv. RELATIVE YIELDS (%)
excess drought date delay date excess drought delay overall
1958 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1959 .0 22.2 125 0. 230 100.0 76.7 100.0 76.7
1960 .0 14.8 125 0. 230 100.0 84.5 100.0 84.5
1961 .0 2.9 125 0. 230 100.0 97.0 100.0 97.0
1962 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1963 .0 44.8 125 0. 230 100.0 52.9 100.0 52.9
1964 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1965 .0 28.5 125 0. 230 100.0 70.1 100.0 70.1
1966 .0 5.7 125 0. 230 100.0 94.0 100.0 94.0
1967 .0 .4 125 0. 230 100.0 99.6 100.0 99.6
1968 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1969 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1970 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1971 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1972 .0 .0 125 0. 230 100.0 100.0 100 0 100.0
1973 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1974 2.2 23.7 125 0. 230 98.5 75.1 100.0 74.0
1975 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1976 .0 .4 125 0. 230 100.0 99.6 100 0 99.6
1977 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1978 .0 37.6 125 0. 230 100.0 60.5 100.0 60.5
1979 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1980 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1981 .0 .2 125 0. 230 100.0 99.8 100.0 99.8
1982 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1983 .0 1.2 125 0. 230 100.0 98.8 100.0 98.8
1984 .0 12.3 125 0. 230 100.0 87.1 100.0 87 1
1985 .0 .0 125 0. 230 100.0 100.0 100 0 100 0
1986 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1987 .0 33.1 125 0. 230 100 0 65.2 100.0 65.2
1988 .0 38.2 125 0. 230 100.0 59.9 100.0 59.9
1989 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1990 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
AVG 1 8 l 125. 0. 230. 100 0 91 5 100.0 91 5

164

165

NB: 0R60 - Drainage Only at 60-Ft Tile Spacings

tti‘ttttttitfitittti‘tittfititfltiittfitttfit*ttt*tittit****t****i*tt*fi *‘ktiiiii‘tiitttt *

 

 

---------- RUN STATISTICS ---------- time: 1/ 7/1993 8 0:20

input file: C:\DN40\1NPUT40\DR6K12.LIS

parameters: free drainage and yields calculated

drain spacing - 1830. cm drain depth - 102.0 cm
SDI - STRESS plant plant harv. RELATIVE YIELDS (X)
excess drought date delay date excess drought delay overall

1958 .0 .0 125 0. 230 100 0 100.0 100.0 100 0
1959 .0 17.0 125 0. 230 100.0 82.2 100.0 82.2
1960 .0 13.0 125 0. 230 100.0 86.3 100.0 86.3
1961 .0 2.0 126 0. 231 100.0 97.9 100.0 97 9
1962 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1963 .0 42.3 125 0. 230 100.0 55.6 100.0 55.6
1964 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1965 .0 26.3 125 0. 230 100.0 72.4 100.0 72.4
1966 .0 4.5 125 0. 230 100.0 95.3 100.0 95.3
1967 .0 .4 125 0. 230 100.0 99.6 100.0 99 6
1968 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1969 6.0 .0 125 0. 230 95.9 100.0 100.0 95.9
1970 .0 .0 125 0. 230 100.0 100 0 100.0 100 0
1971 .0 .0 125 0. 230 100.0 100.0 100 0 100 0
1972 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1973 .0 .0 125 0. 230 100.0 100 0 100.0 100 0
1974 13.1 15.7 125 0. 230 91.1 83.5 100.0 76.1
1975 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1976 3.0 .0 125 0. 230 98.0 100.0 100.0 98 0
1977 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1978 .0 35.2 125 0. 230 100.0 63.1 100.0 63 1
1979 .0 .0 125 0. 230 100 0 100.0 100.0 100 0
1980 .0 .0 125 0. 230 100 0 100.0 100.0 100.0
1981 .0 .0 125 0. 230 . 100.0 100.0 100.0 100.0
1982 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1983 .0 .0 125 0. 230 100 0 99.9 100.0 99.9
1984 .0 9.3 125 0. 230 100.0 90.2 100.0 90.2
1985 .0 .0 125 0. 230 100 0 100.0 100.0 100 0
1986 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1987 .0 31.2 125 0. 230 100.0 67.2 100.0 67 2
1988 .0 36.6 125 0. 230 100.0 61.5 100 0 61 5
1989 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1990 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
AVE 7 7 1 125. 0. 230. 99.5 92.6 100.0 92.2

166

N4: S120 - Subirrigation at 20-Ft Tile Spacings

t‘ktttt*****”***i*tti’tfl‘itttttitfitt*tt‘ktitttt‘ktttttttttittttttttt‘kttttttttti it t t t

---------- RUN STATISTICS ---------- time: 1/ 7/1993 8 0:34
input file: C:\DM40\1NPUT40\SIZK12.LIS
parameters: subirrigation run and yields calculated

drain spacing - 610. cm drain depth - 102.0 cm

 

 

 

SDI - STRESS plant plant harv. RELATIVE YIELDS (%)
excess drought date delay date excess drought delay overall
1958 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1959 9.1 .0 125 0. 230 93.8 100.0 100 0 93.8
1960 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1961 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1962 .1 .0 125 0. 230 99.9 100.0 100.0 99.9
1963 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1964 1.4 .0 125 0. 230 99.1 100 0 100.0 99.1
1965 .0 .0 125 0. 230 100 0 100.0 100.0 100.0
1966 .0 .0 125 0. 230 100.0 100.0 100 0 100.0
1967 2.9 .0 125 0. 230 98.0 100.0 100.0 98.0
1968 26.5 .0 125 0. 230 82.0 100.0 100.0 82.0
1969 11.5 .0 125 0. 230 92.2 100.0 100.0 92.2
1970 7.4 .0 125 0. 230 95.0 100.0 100.0 95.0
1971 .l .0 125 0. 230 99.9 100.0 100.0 99.9
1972 12.5 .0 125 0. 230 91.5 100.0 100.0 91.5
1973 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1974 1.4 .0 125 0. 230 99.0 100.0 100 0 99 0
1975 3.7 .0 125 0. 230 97.5 100.0 100.0 97.5
1976 21.2 .0 125 0. 230 85.6 100.0 100 0 85.6
1977 .1 .0 125 0. 230 99.9 100.0 100.0 99.9
1978 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1979 12.6 .0 125 0. 230 91.5 100.0 100.0 91.5
1980 10.7 .0 125 0. 230 92.7 100 0 100.0 92.7
1981 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1982 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1983 2.9 .0 125 0. 230 98.0 100.0 100.0 98.0
1984 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1985 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1986 .1 .0 125 0. 230 100.0 100.0 100.0 100.0
1987 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1988 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1989 .5 .0 125 0. 230 99.7 100.0 100.0 99.7
1990 .0 .0 125 0. 230 100 0 100 0 100.0 100.0
AVG 3.8 0 125. 0. 230. 97.4 100 0 100.0 97.4

167

H5: 5130 - Subirrigation at 30-Ft Tile Spacings

*ititi’itiittit‘kti*i’fi‘ktfiit**********itifiiitiittt‘tiifiti’ttt*iti’i‘itttttitttttifi*i'ttt

---------- RUN STATISTICS ---------- time: 1/ 7/1993 9 0:43
input file: C:\DM40\INPUT40\SI3K12.LIS
parameters: subirrigation run and yields calculated

drain spacing - 914. cm drain depth - 102.0 cm

 

 

 

S01 - STRESS plant plant harv. RELATIVE YIELDS (X)
excess drought date delay date excess drought delay overall
1958 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1959 5.9 .0 125 0. 230 96.0 100.0 100.0 96.0
1960 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1961 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1962 .5 .0 125 0. 230 99.7 100.0 100.0 99.7
1963 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1964 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1965 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1966 .0 .O 125 O. 230 100.0 100.0 100.0 100.0
1967 1.7 .1 125 0. 230 98.8 99.9 100.0 98.8
1968 11.9 .0 125 0. 230 91.9 100.0 100 0 91.9
1969 8.3 .0 125 0. 230 94.3 100.0 100.0 94 3
1970 7.3 .0 125 0. 230 95.0 100.0 100.0 95.0
1971 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1972 10.2 .0 125 0. 230 93.1 100.0 100.0 93.1
1973 .O .0 125 0. 230 100.0 100.0 100.0 100 0
1974 4.9 .0 125 0. 230 96.7 100 0 100.0 96.7
1975 .2 .O 125 0. 230 99.9 100.0 100.0 99.9
1976 2.6 .0 125 0. 230 98.2 100.0 100.0 98.2
1977 .0 .0 125 0. 230 100 0 100.0 100.0 100.0
1978 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1979 .7 .0 125 0. 230 99.5 100.0 100.0 99.5
1980 3.4 .0 125 0. 230 97.7 100.0 100.0 97.7
1981 .0 .O 125 O. 230 100.0 100.0 100.0 100.0
1982 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1983 .7 .0 125 0. 230 99.6 100.0 100.0 99.6
1984 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1985 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1986 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1987 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1988 .0 .0 125 0. 230 100 0 100.0 100 0 100.0
1989 .4 .0 125 0. 230 99.7 100.0 100.0 99.7
1990 .0 .0 125 0. 230 100.0 100.0 100 0 100.0
AVG 1 8 0 125. 0. 230. 98.8 100.0 100.0 98.8

168

HS: 8160 - Subirrigation at 60-Ft Tile Spacings

*ttiitttiititi’ii’i’tittt*tittttit‘kttttttttttittti’tfiitttitt*tttiitttttttttttttttt it

---------- RUN STATISTICS ---------- time: 1/ 7/1993 0 0:38
input file: C:\DM40\INPUT40\516K12.LIS
parameters: subirrigation run and yields calculated

drain spacing - 1830. cm drain depth . 102.0 cm

 

 

SDI - STRESS plant plant harv. RELATIVE YIELDS (%)
excess drought date delay date excess drought delay overall
1958 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1959 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1960 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1961 .0 .0 126 0. 231 100.0 100.0 100.0 100.0
1962 .5 .0 125 0. 230 99.7 100.0 100.0 99.7
1963 .0 13.2 125 0. 230 100.0 86.1 100.0 86 1
1964 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1965 .0 3.0 125 0. 230 100.0 96.9 100.0 96.9
1966 .0 3.0 125 0. 230 100.0 96.8 100.0 96.8
1967 .0 .3 125 0. 230 100 O 99.7 100.0 99.7
1968 3.7 .0 125 0. 230 97.5 100.0 100.0 97 5
1969 6.0 .0 125 0. 230 95.9 100.0 100.0 95 9
1970 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1971 .0 .0 125 0. 230 100 0 100.0 100 0 100.0
1972 .3 .0 125 0. 230 99.8 100.0 100 0 99.8
1973 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1974 14.4 .0 125 0. 230 90.2 100.0 100.0 90 2
1975 .0 .0 125 0. 230 100.0 100.0 100 0 100 0
1976 3.0 .0 125 0. 230 98.0 100.0 100 0 98.0
1977 .0 .0 125 0. 230 100.0 100.0 100 0 100 0
1978 .0 5.0 125 0. 230 100.0 94.8 100.0 94 8
1979 .0 .0 125 O. 230 100.0 100.0 100.0 100 O
1980 .0 .0 125 0. 230 100.0 100 0 100.0 100 0
1981 .0 .0 125 0. 230 100 0 100.0 100.0 100 0
1982 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1983 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
1984 .0 .0 125 O. 230 100.0 100.0 100.0 100 0
1985 .0 .0 125 0. 230 100 0 100.0 100 0 100.0
1986 .0 .0 125 0. 230 100.0 100 0 100.0 100.0
1987 .0 29.9 125 0. 230 100.0 68.6 100.0 68.6
1988 .0 33.4 125 0. 230 100.0 64.9 100 0 64.9
1989 .0 .0 125 0. 230 100.0 100.0 100.0 100.0
1990 .0 .0 125 0. 230 100.0 100.0 100.0 100 0
AVG 8 2 7 125. 0. 230. 99.4 97.2 100.0 96.6

APPENDIX 1

SIMULATION HATER BALANCE RESULTS

11: DR20 - Drainage Only at 20-Ft Tile Spacings

*ifitt***t****i****tt**t‘ti‘kfifitfi‘ti’ttfifiiifii‘t."*tt‘kttittt ** *t*fi**********t**t * t *‘k t

---------- RUN STATISTICS ---------- time: 1/ 7/1993 8 0:25
input file: C:\0M40\1NPUT40\DR2K12.LIS
parameters: free drainage and yields calculated

drain spacing - 610. cm drain depth 8 102.0 cm

 

YEAR RAINFALL INFILTR ET DRAIN RUNOFF DRYDAYS UORKDAYS SEH PUMP VOL

1958 51.00 51.00 44.91 11.91 .00 28.00 98.80 .0 .00
1959 95.83 95.83 56.58 38.41 .00 31.00 87.28 .0 .00
1960 56.97 56.97 40.54 18.54 .00 46.00 98.00 .0 .00
1961 77.57 77.57 52.37 24.01 .00 27.00 93.25 .0 .00
1962 59.16 59.16 39.83 19.45 .00 52.00 87.83 .0 .00
1963 45.92 45.92 35.60 10.20 .00 56.00 98.00 .0 .00
1964 62.46 62.46 43.87 18.54 .00 37.00 95.17 .0 .00
1965 65.51 65.51 40.20 25.12 .00 53.00 100.03 .0 .00
1966 52.17 52.17 35.01 17.23 .00 61.00 99.20 .0 .00
1967 76.94 76.94 42.91 33.97 .00 40.00 95.90 .0 .00
1968 86.11 86.11 53.71 31.66 .00 13.00 100.58 .0 .00
1969 72.59 72.59 43.84 29.83 .00 13.00 94.92 .0 .00
1970 75.39 75.39 51.32 24.00 .00 26.00 91.40 .0 .00
1971 67.36 67.36 42.35 23.21 .00 34.00 97.47 .0 .00
1972 96.60 96.60 55.29 40.94 .00 14.00 81.55 .0 .00
1973 84.79 84.79 44.26 42.04 .00 42.00 91.42 .0 .00
1974 82.27 82.27 38.19 43.99 .00 49.00 95.90 .3 .00
1975 115.27 115.27 56.09 59.37 .00 26.00 83.61 .0 .00
1976 81.69 81.69 44.44 37.69 .00 40.00 88.87 .0 .00
1977 73.30 73.30 51.55 21.76 .00 42.00 88.75 .0 .00
1978 60.17 60.17 39.56 17.75 .00 60.00 91.97 .0 .00
1979 64.06 64.06 43.97 22.61 .00 23.00 105.00 .0 .00
1980 80.98 80.98 54.51 25.90 .00 18.00 90.85 .0 .00
1981 86.51 86.51 53.65 33.20 .00 35.00 86.48 .0 .00
1982 72.34 72.34 46.06 26.54 .00 37.00 98.70 .0 .00
1983 81.64 81.64 50.99 30.83 .00 33.00 81.50 .0 .00
1984 76.58 76.58 46.76 27.77 .00 46.00 88.28 .0 .00
1985 103.17 103.17 52.54 52.72 .00 36.00 85.40 .0 .00
1986 95.43 95.43 57.33 38.26 .00 23.00 84.98 .0 .00
1987 74.27 74.27 51.53 22.63 .00 69.00 92.38 .0 .00
1988 70.21 70.21 43.52 26.24 .00 74.00 98.00 .0 .00
1989 82.14 82.14 55.55 25.95 .00 20.00 90.53 .0 .00
1990 84.02 84.02 47.68 36.24 .00 43.00 91.33 .0 .00
AVG 76.07 76.07 47.17 29.05 .00 37.79 92.53 0 .00

169

12: DR30 - Drainage Only at 30-Ft Tile Spacings

ti‘fii’i***itt***ii**********tit‘ktifiiittttttitittttttiittttttttttttitttttt*ttttt*tt

---------- RUN STATISTICS ---------- time: 1/ 7/1993 9 0:29
input file: C:\DM40\INPUT40\DR3K12.LIS
parameters: free drainage and yields calculated

drain spacing - 914. cm drain depth - 102.0 cm

 

 

YEAR RAINFALL INFILTR ET DRAIN RUNOFF DRYDAYS HORKDAYS SEH PUMP VOL

1958 51.00 51.00 45.08 11.49 .00 27.00 98.80 .0 .00
1959 95.83 95.83 57.06 37.34 .00 30.00 87.28 .0 .00
1960 56.97 56.97 40.73 19.04 .00 46.00 98.00 .0 .00
1961 77.57 77.57 52.56 23.60 .00 27.00 93.25 .0 .00
1962 59.16 59.16 40.22 19.21 .00 48.00 87.83 .0 .00
1963 45.92 45.92 35.84 9.96 .00 56.00 97.00 .0 .00
1964 62.46 62.46 44.10 18.13 .00 35.00 95.17 .0 .00
1965 65.51 65.51 40.47 24.46 .00 53.00 100.03 .0 .00
1966 52.17 52.17 35.12 17.47 .00 61.00 99.20 .0 .00
1967 76.94 76.94 43.31 33.34 .00 40.00 95.90 .0 .00
1968 86.11 86.11 53.94 31.14 .00 13.00 100.58 .0 .00
1969 72.59 72.59 43.98 30.39 .00 11.00 94.92 .0 .00
1970 75.39 75.39 51.63 23.57 .00 26.00 91.40 .0 .00
1971 67.36 67.36 42.48 22.52 .00 33.00 97.47 .0 .00
1972 96.60 96.60 55.35 40.96 .00 14.00 81.55 .0 .00
1973 84.79 84.79 44.48 41.82 .00 41.00 91.42 .0 .00
1974 82.27 82.27 38.77 43.78 .00 47.00 95.90 11.1 .00
1975 115.27 115.27 56.25 59.07 .00 26.00 83.14 .0 .00
1976 81.69 81.69 44.76 37.61 .00 39.00 88.33 .0 .00
1977 73.30 73.30 51.81 21.39 .00 42.00 88.75 .0 .00
1978 60.17 60.17 39.78 17.46 .00 60.00 91.97 .0 .00
1979 64.06 64.06 44.15 22.19 .00 23.00 105.00 .0 .00
1980 80.98 80.98 55.08 25.44 .00 18.00 90.85 .0 .00
1981 86.51 86.51 53.88 33.18 .00 35.00 86.48 .0 .00
1982 72.34 72.34 46.39 26.02 .00 35.00 98.70 .0 .00
1983 81.64 81.64 51.23 30.74 .00 33.00 81.50 .0 .00
1984 76.58 76.58 47.11 26.87 .00 46.00 88.28 .0 .00
1985 103.17 103.17 52.75 53.12 .00 35.00 85.36 .0 .00
1986 95.43 95.43 57.55 38.14 .00 23.00 84.18 .0 .00
1987 74.27 74.27 51.70 22.22 .00 69.00 92.38 .0 .00
1988 70.21 70.21 43.63 26.07 .00 74.00 98.00 .0 .00
1989 82.14 82.14 55.76 25.98 .00 20.00 90.53 .0 .00
1990 84.02 84.02 48.17 35.23 .00 42.00 91.33 .0 .00
AVG 76.07 76.07 47.43 28.76 .00 37.21 92.44 3 .00

170

13: DR60 - Drainage Only at 60—Ft Tile Spacings

******fittiittitittfifliiiitttt*ttiitttiiititfifitttititt*iitittttttitttfitttitt*‘A’tttt

---------- RUN STATISTICS ---------- time: 1/ 7/1993 0 0:20
input file: C:\DN40\INPUT40\DR6K12.LIS
parameters: free drainage and yields calculated

drain spacing - 1830. cm drain depth - 102.0 cm

 

YEAR RAINFALL INFILTR ET DRAIN RUNOFF DRYDAYS UORKDAYS SEV PUMP VOL

1958 51.00 51.00 46.09 10.32 .00 26.00 98.80 .0 .00
1959 95.83 95.83 59.27 33.54 .00 25.00 87.28 .0 .00
1960 56.97 56.97 41.92 19.73 .00 41.00 98.00 .0 .00
1961 77.57 77.57 53.52 21.54 .00 23.00 90.75 .0 .00
1962 59.16 59.16 42.28 17.57 .00 41.00 87.83 .0 .00
1963 45.92 45.92 37.57 8.79 .00 48.00 96.30 .0 .00
1964 62.46 62.46 45.71 15.77 .00 31.00 95.17 .0 .00
1965 65.51 65.51 41.48 22.35 .00 52.00 100.03 .0 .00
1966 52.17 52.17 36.01 17.22 .00 56.00 99.20 .0 .00
1967 76.94 76.94 45.05 30.98 .00 35.00 93.43 .0 .00
1968 86.11 86.11 54.68 30.13 .00 13.00 100.58 .0 .00
1969 72.59 72.59 44.81 30.74 .00 7.00 91.12 30.1 .00
1970 75.39 75.39 53.21 21.69 .00 26.00 91.40 .0 .00
1971 67.36 67.36 43.52 20.49 .00 32.00 96.47 .0 .00
1972 96.60 96.60 55.73 40.63 .00 11.00 76.85 .0 .00
1973 84.79 84.79 46.38 39.94 .00 37.00 91.42 .0 .00
1974 82.27 82.27 41.36 41.89 .00 38.00 93.72 65.7 .00
1975 115.27 115.27 57.18 57.64 .00 23.00 74.29 .0 .00
1976 81.69 81.69 46.39 36.84 .00 27.00 82.11 15.0 .00
1977 73.30 73.30 53.42 19.49 .00 37.00 88.75 .0 .00
1978 60.17 60.17 40.84 16.60 .00 58.00 91.97 .0 .00
1979 64.06 64.06 45.08 20.49 .00 21.00 105.00 .0 .00
1980 80.98 80.98 56.14 24.77 .00 17.00 90.85 .0 .00
1981 86.51 86.51 55.44 32.02 .00 27.00 82.12 .0 .00
1982 72.34 72.34 47.64 24.22 .00 34.00 98.70 .0 .00
1983 81.64 81.64 53.11 29.10 .00 27.00 80.50 .0 .00
1984 76.58 76.58 48.92 24.27 .00 41.00 88.28 .0 .00
1985 103.17 103.17 54.10 52.73 .00 34.00 83.18 .0 .00
1986 95.43 95.35 58.71 37.03 .07 21.00 79.39 .0 .00
1987 74.27 74.27 52.74 20.56 .00 69.00 92.38 .0 .00
1988 70.21 70.21 44.95 25.14 .00 72.00 98.00 .0 .00
1989 82.14 82.14 56.58 25.52 .00 19.00 90.53 .0 .00
1990 84.02 84.02 50.09 32.44 .00 30.00 91.33 .0 .00
AVG 76.07 76.07 48.79 27.34 .00 33.30 91.08 3 4 .00

171

 

14: $120 - Subirrigation at 20-Ft Tile Spacings

*ttttfiitti.*itittfitttititttttttii‘t*ttttttttitttttttttttt *ttt‘ttf‘kitttttt *ttt * t** t

---------- RUN STATISTICS ---------- time: 1/ 7/1993 9 0:34
input file: C:\DM40\INPUT40\SIZK12.LIS
parameters: subirrigation run and yields calculated

drain spacing - 610. cm drain depth - 102.0 cm

 

YEAR RAINFALL INFILTR ET DRAIN RUNOFF DRYDAYS HORKDAYS SEH PUMP VOL
4

1958 51.00 51.00 52.77 .05 .00 21.00 98.80 .0 -16.92
1959 95.83 95.83 62.30 32.70 .00 12.00 87.28 72.3 -20.78
1960 56.97 56.97 52.78 5.32 .00 14.00 98.00 .0 -20.71
1961 77.57 77.57 56.15 21.21 .00 13.00 93.25 .0 -19.15
1962 59.16 59.16 52.40 6.88 .00 26.00 87.33 .5 ~18.37
1963 45.92 45.92 55.35 -9.59 .00 5.00 96.30 .O -26.31
1964 62.46 62.46 55.30 7.15 .00 14.00 95.17 4.8 -19.06
1965 65.51 65.51 56.21 9.11 .00 22.00 99.03 .0 -24.36
1966 52.17 52.17 53.51 -1.28 .00 11.00 99.20 .0 -24.95
1967 76.94 76.94 55.03 21.86 .00 16.00 93.30 9.0 -19.25
1968 86 11 86 11 54 74 30 63 00 13 00 100 58 112.6 -17.06
1969 72 59 72 59 48 02 25 65 00 4 00 94 92 41.8 -14.32
1970 75 39 75 39 58.42 16 9O 00 14 DO 91 40 26.4 -16.91
1971 67 36 67 36 54.95 10 61 OO 10 OO 96 47 .7 -19.92
1972 96 60 96 60 55.74 40 48 OO 14 00 81 55 65.8 -15.12
1973 84 79 84 79 58.94 27 36 00 5 00 91 42 .0 -20.46
1974 82 27 82 27 54.53 27 64 OO 8 00 95 90 7.2 -19.33
1975 115 27 113 14 60.99 52 35 2 13 8 00 81 51 11.5 -19.27
1976 81 69 81 69 58.03 24 10 00 3 00 87 57 66.2 -20.57
1977 73 30 73 30 58.94 14 37 00 29 00 88 75 .2 -18.97
1978 60.17 60.17 59.03 -1 72 .00 12.00 91.97 .0 -24.06
1979 64.06 64.06 51.18 15.40 .00 12.00 105.00 39.7 -19.89
1980 80 98 80 98 55.77 24 64 00 18 00 90 85 45.2 -15.57
1981 86 51 86 51 58.81 28 03 00 12 00 86 48 .0 -18.40
1982 72 34 72 34 56.80 15 80 00 19 00 98 7D .0 -18.67
1983 81 64 81 64 63.89 17 93 00 7 00 80 50 9.1 -20.69
1984 76 58 76 58 60.96 13 57 00 14 00 87 28 .0 -22.00
1985 103 17 103 17 58.46 46 81 OO 15 00 83 32 .0 -19.79
1986 95 43 95 43 59.01 36 58 00 14 00 84 88 .2 -17.24
1987 74 27 74 27 65.71 8 45 00 20 00 92 38 .0 -27.73
1988 70 21 70 21 62.19 7 57 OO 31 OO 98 00 .0 -27.69
1989 82 14 82 14 55.94 25 55 OO 20 00 9O 53 6.0 -16.25
1990 84 02 84 02 60.71 23 20 00 11 OO 91 33 .0 -18.17
AVG 76.07 76.01 57.08 19.07 .06 14.15 92.09 15.7 -19.94

172

 

15: $130 - Subirrigation at 30-Ft Tile Spacings

"tiiiitifiit‘kififi'i‘tittiit*ii‘tiifitit*tfit*‘titfittifiiii*ii***i*t***********iiit*tt*tt

---------- RUN STATISTICS ---------- time: 1/ 7/1993 9 0:43
input file: C:\DM40\INPUT40\SI3K12.LIS
parameters: subirrigation run and yields calculated

drain spacing - 914. cm drain depth a 102.0 cm

 

YEAR RAINFALL INFILTR ET DRAIN RUNOFF DRYDAYS HORKDAYS SEN PUMP VOL

1958 51.00 51.00 52.88 3.68 .00 21.00 98.80 .0 -14.40
1959 95.83 95.83 62.40 32.13 .00 11.00 87.28 48.2 -17.70
1960 56.97 56.97 52.98 5.65 .00 15.00 98.00 .0 -19.24
1961 77.57 77.57 56.37 20.88 .00 13.00 93.25 .0 -16.60
1962 59.16 59.16 52.02 7.37 .00 24.00 87.33 1.8 -14.97
1963 45.92 45.92 55.41 -9.73 .00 7.00 96.30 .0 -23.81
1964 62.46 62.46 54.76 7.58 .00 14.00 95.17 .0 -15.46
1965 65.51 65.51 56.25 8.76 .00 22.00 99.03 .0 -22.42
1966 52.17 52.17 53.01 -.46 .00 11.00 99.20 .0 -22.36
1967 76.94 76.94 54.36 22.31 .00 19.00 93.30 5.4 ~16.14
1968 86.11 86.11 55.06 30.04 .00 13.00 100.58 68.7 -13.51
1969 72.59 72.59 47.92 26.37 .00 4.00 94.92 30.5 -10.50
1970 75.39 75.39 58.54 16.68 .00 14.00 91.40 26.0 -13.78
1971 67.36 67.36 54.77 10.29 .00 10.00 96.47 .D -16.45
1972 96.60 96.60 55.69 40.62 .00 14.00 81.55 53.5 -12.27
1973 84.79 84.79 58.58 27.72 .00 5.00 91.42 .0 -l7.76
1974 82.27 82.27 54.87 27.64 .00 7.00 95.76 24.5 -17.74
1975 115.27 113.35 60.99 52.44 1.92 8.00 81.14 .6 -15.70
1976 81.69 81.69 57.70 24.65 .00 3.00 87.13 8.1 -17.65
1977 73.30 73.30 58.93 14.30 .00 29.00 88 75 .0 -15.42
1978 60.17 60.17 58.10 -.87 .00 12.00 91.97 .D -22.08
1979 64.06 64.06 51.25 15.13 .00 12.00 105.00 2.1 -16.60
1980 80.98 80.98 55.87 24.63 .00 17.00 90.85 14.0 -12.39
1981 86.51 86.51 58.85 28.19 .00 12.00 86.23 .0 -I4.95
1982 72.34 72.34 56.80 15.64 .00 20.00 98.70 .0 -15.94
1983 81.64 81.64 64.14 17.81 .00 7.00 80.50 2.1 -l7.09
1984 76.58 76.58 61.09 12.94 .00 14.00 87.28 .0 ‘18.80
1985 103.17 103.17 58.57 47.23 .00 15.00 82.84 .0 ~16.93
1986 95 43 95 43 59.09 36 58 OO 14 OO 84 05 .O -13.83
1987 74 27 74 27 65.66 8 30 OO 20 00 92 38 .0 -25.53
1988 7D 21 70 21 61.99 7 70 OO 31 OO 98 00 .0 -25.41
1989 82 14 82 14 56.12 25 59 00 20 00 90 53 4.9 -12.07
1990 84 02 84 02 60.80 22 65 00 7 00 91 33 .0 -15.66
AVG 76 O7 76 01 57 02 19 11 06 14 09 92 01 8 8 -17.01

173

16: 5160 - Subirrigation at 60-Ft Tile Spacings

*tiitti'i'*ttitti’i’fitttitti‘itit*‘kt‘l’fi*tfi‘ttttt*tti‘tifitti’iitttitittfi’ttiittt*tt*ttt

---------- RUN STATISTICS ---------- time: 1/ 7/1993 9 0:38
input file: C:\DM40\INPUT40\S16K12.LIS
parameters: subirrigation run and yields calculated

drain spacing - 1830. cm drain depth - 102.0 cm

 

YEAR RAINFALL INFILTR ET DRAIN RUNOFF DRYDAYS HORKDAYS SEH PUMP VOL

1958 51.00 51.00 51.95 3.90 .00 21.00 98.80 .0 -7.83
1959 95.83 95.83 63.12 30.25 .00 10.00 87.12 .0 -9.31
1960 56.97 56.97 50.82 9.20 .00 13.00 98.00 .D -11.10
1961 77.57 77.57 57.54 19.07 .00 11.00 90.75 .0 -10.17
1962 59.16 59.16 49.60 10.22 .00 20.00 87.33 1.8 -7.27
1963 45.92 45.92 46.13 -.56 .00 29.00 96.30 .0 -9.89
1964 62.46 62.46 52.76 9.22 .00 13.00 95.17 .0 -8.06
1965 65.51 65.51 50.44 13.69 .00 30.00 100.03 .0 -11.03
1966 52.17 52.17 44.97 8.27 .00 30.00 99.20 .0 -10.36
1967 76.94 76.94 52.17 23.93 .00 22.00 90.83 .0 -8.54
1968 86.11 86.11 56.35 28.46 .00 13.00 100.58 43.0 -6.17
1969 72.59 72.59 47.92 27.61 .00 1.00 90.99 30.2 -4.24
1970 75.39 75.39 58.90 16.02 .00 14.00 91.40 .0 -7.37
1971 67.36 67.36 52.07 11.89 .00 10.00 96.47 .0 -IO.4O
1972 96.60 96.60 56.02 40.40 .00 11.00 74.15 1.3 -6.86
1973 84.79 84.79 54.69 31.64 .00 9.00 91.42 .D -9.38
1974 82.27 82.27 50.84 32.40 .00 12.00 92.83 72.1 -9.97
1975 115.27 114.89 61.22 53.25 .37 8.00 73.89 .0 -7.35
1976 81.69 81.69 54.74 28.46 .00 9.00 80.81 15.0 -8.85
1977 73.30 73.30 59.32 13.62 .00 27.00 88.75 .0 -9.38
1978 60.17 60.17 49.88 7.46 .00 41.00 91.97 .0 -10.61
1979 64.06 64.06 50.67 14.95 .00 12.00 105.00 .0 -9.38
1980 80.98 80.98 56.53 24.43 .00 17.00 90.85 .0 -7.00
1981 86.51 86.51 59.45 28.03 .00 7.00 79.76 .0 -9.02
1982 72 34 72 34 55.04 16 81 DO 20 DO 98 7O .0 -9.33
1983 81 64 81 64 61.44 20 76 00 8 00 80 50 .0 -9.13
1984 76 58 76 58 57.92 15 26 00 15 00 85 62 .0 -10.02
1985 103 17 103 17 59.45 47 39 00 15 00 80 50 .O -10.20
1986 95 43 94 65 59.79 35 25 78 14 00 76 83 .O -7.44
1987 74 27 74 27 57 67 15 64 DO 62 DO 92 38 .0, -7 38
1988 70 21 70 21 47.43 22 66 00 65 00 98 00 .0 -3.80
1989 82 14 82 14 56.92 25 17 OO 19 00 89 88 .0 -5.46
1990 84 02 84 02 58.64 23 89 00 8 00 90 25 .0 -9.14
AVG 76 O7 76 D4 54 62 21.47 03 18 67 90 46 5 O -8 53

174

BIBLIOGRAPHY

Anderson, J. R. 1974a. ”Simulation: Methodology and Application 1n Agricultural
Economics" WW 42(1) 3 55

Anderson, J. R. 1974b. ”Risk Efficiency 1n the Interpretation of Agricultural Production
Research" MmeiMarkstiauadAsn’eulflnLEcgmm 131- 184

Anderson, J.R., J .L. Dillon, and B. Hardaker. 1977. Agricultural Decision Malysis. Ames
Iowa: Iowa State University Press, 1977.

Anderson. R.L. 1968. "A Simulation Program to Establish Optimum Crop Patterns on
Irrigated Farms Based on Preseason Estimates of Water Supply.” Amefign

lggmfl of Agricultural mnomics. 50(5): 1586-1590.

Anderson, R.L. and A. Maas. 1971. "A Simulation of Irrigation Systems." Technical
Bulletin 1431. Washington, DC: USDA-ERS.

Antle, J .M. 1982. ”Incorporating Risk in Production Analysis. W
W 66:1099-1106.

Auernhamer, D. and H.W. Belcher. 1990. "Alfalfa Production with Subirrigation". In F.M.
D’Itri and J.A. Kubitz, eds. e Sa ° aw Ba M'chi an Sub' ' at' Draina e
21mg; 1987-88. East Lansing: Institute of Water Research.

Auernhamer, M. E. 1991. "Apportionment of Benefits within an Irrigation District. In

F. M. D’Itri and J .A Kubitz, eds. W
P 0 989-9 .East Lansing: Institute of Water

Research.

Barry, P.J., J.A. Hopkin, and CB. Baker. 1979. W.
2nd ed., Danville, 111.: Interstate Publishers.

Batra, R. N. and A. Ullah. 1974. "Competitive Firm and the Theory of Input Demand
under Price Uncertainty." MW. 40: 537-548.

Belcher, H. W. 1990a. "Regional Approach to Subirrigation Planning.” In F. M. D’Itri and
J.A. KUbImCdS. l1- " i 1 , , _,,,,'.° 0
1282-88. East Lansing: Institute of water Research.

 

175

 

176

Belcher, H..W 1990b. W t abe ' ' no 'c cien
W. A Dissertation. Michigan State University. Department off
Agricultural Engineering. East Lansing, Michigan.

Belcher, H. W. and Erica Wood. 1990. “Saginaw Bay 1988 Subirrigation Inventory". In

F M D’Itri and IA. KubiIZ. eds W
W. East Lansing: Institute of Water

Research.

Black, J. R. and S. R. Thompson. 1978. ”Some Evidence on Weather-Crop-Yield
Interaction" WW ??( ?) 540-543

Blackie, MJ. and KC. Schneeberger. 1971. "An Irrigation Investment Simulation.”
Bhggesiag Journal of Economics. 5(3): 11-15.

Boggess, W.G., J .W. Jones, D. P. Swaney, and GD. Lynne. 1981. "Evaluating Irrigation

Strategies in Soybeans: A Simulation Approach.” lm'gatigg Scheduling E9; Water

WW. Proceedings of the American Society of
Agricultural Engineers Irrigation Scheduling Conference. ASAE Publication 23-

81.

Boggess, W.G., C.B. Amerling, and GD. Lynne. 1982. ”Evaluating Irrigation Investment
Decisions." In: oc. o t e ecial Conference 0 nviro mentall Sou

WWW Irrigation and
Drainage Division of the ASCE. New York: American Society of Civil Engineers.

Boggess, W.G., and CB. Amerling. 1983. "A Bioeconomic Simulation Analysis of
Irrigation Investments.” WWW- 15(2)= 35-92-

Boggess, W. G., G. D. Lynne, J. W. Jones, and D. P. Swaney. 1983. "Risk-Return
Assessment of Irrigation Decisions 1n Humid Regions. Sggthem lggmal of

WW 15(1) 135-144

Brink, L and B. McCarl. 1978. “The Tradeoff between Expected Return and Risk among
Corn Belt FamerS" WW 60: 259-263

Burt, O. R. and R. M. Finley. 1971. ”Economic Analysis of Irrigation 1n Subhumid
ClimateS" WW 53: 3346

Campbell, K.L., J. S. Rogers, and D. R. Hensel. 1985. "Drainage Water Quality from

Potato Production.” W. 28(6): 1798- 1801. ASAE: St.
Joseph, MI.

Carruthers, ID. and G.F. Donaldson. (1971) "Estimation of Effective Risk Reduction

through Irrigation of a Perennial Crop." Wig. 12(1):
39-48.

177

Cassidy, P.A. J. L. Rodgers and W. O. McCarthy. 1970. "A Simulation Approach to Risk

Assessment on Investment 1111488"me
Economics 38(1) 3-24

Charlton, PJ. and S. C. Thompson. 1970. "Simulation of Agricultural Systems." M
We 21(3) 373-334

Clouser, R., W.L. Miller, P. Erhabor, and J. Sobek. 1981. 'I‘he Economic Opportunity to
Irrigate Corn and Soybeans on Midwestern Claypan Soils." Purdue University
Water Resources Center Technical Report Number 93.

Cochran MJ. and R Raskin 1988 MW
a 0 .Staff Paper SP0688.
Department of Agricultural Economics Vand Rural Sociology, University of
Arkansas, Fayetteville, AR.

D’Itri, F.M. and J.A. Kubitz, eds. 1991. W

b' ' atio ' o' ' 989- .East Lansing: Institute of Water
Research.

D’Itri, F. M. and J.A. Kubitz, eds. 19%.f1he Saginaw Bay, Mic chigan
W. 8East Lansing: Institute of Water

Research.

Day, RH, 1965. "Probability Distributions of Field Crop Yields." MALQLEM
M. 47:713-741.

DeBoer, D. W., D. D. Brosz, and 1.1. Wiersma. 1977. "Irrigation Application Depths for
Optimum Crop Production. W. 20: 1067-78.

El-Saadi, MA. A. and ER. Swanson. 1972. "Effect of Supplemental Irrigation on Income

and Net Worth: A Simulation Analysis." WW. 12(1): 1-
5.

Evans, R. O. and R.W. Skaggs. 1988. "Design Procedures for Water Table Management
Systems in North Carolina". Proceedings Third International Workshop on Land
Drainage. Ohio State University, Columbus, Ohio. pp. D9-D22.

Evans, R. O., R.W. Skaggs, and RE. Sneed. 1988. "Economics of Controlled Drainage
and Subirrigation Systems". Bulletin 397. Raleigh, NC: Agricultural Extension
Service. North Carolina State University.

Evans, R. O., R.W. Skaggs, and RE. Sneed. 1990. "Normalized Crop Susceptibility
Factors for Corn and Soybean to Excess Water Stress". MOW
ASAE. 33(4):1153-1161.

 

178

Fogiel, A. and H.W. Belcher. 1991. "Research Literature Review of Water Table
Management Impacts on Water Quality". East Lansing, Michigan: Agricultural
Engineering Department. Michigan State University.

Fogiel, AC. 1992. "Water Quality and Biomass Impacts of Water Table Management." A
Thesis. Michigan State University. Department of Agricultural Engineering. East
Lansing, MI.

Friedman, M. and L. H. Savage. 1948. 'Ihe Utility Analysis of Choices Involving Risk" "
W 56: 279-304

Fulton, J. R., R. P. King, and P. L. Fackler. 1988. thining Farma and Coung Data to

W Staff Paper Series P8840.
Department of Agricultural and Applied Economics. University of Minnesota.

Institute of Agriculture, Forestry and Home Economics: St. Paul, MN.

Gowon, D.T., J .C. Andersen, and B. Biswas. 1978. "An Economic Interpretation of the
Impact of Phenologically Timed Irrigation on Corn Yield." Wgstern Jgurnal of

Agriculturg Eggnomics. 3: 145-156.
Gujarati, D. 1978. Basig mnometrics. New York: McGraw-Hill.

Hadar, J. and W.R. Russell. 1969. "Rules for Ordering Uncertain Prospects." mm
W 59(1) 25-34-

Hall, WA. and W5. Butcher. 1968. "Optimum Irrigation Practice Under Conditions of
Deficient Water Supply." Transactions of the ASAE. 4: 131-134.

Halter, AN. and G.W. Dean. 1965. "Use of Simulation in Evaluating Management
Strategies under Uncertainty." =lgumal of Farm Economics. 47(3): 557-573.

Hammersley, J.M. and DC. Hanscomb. 1964. Monte Carlg Mgthods. London: Metheun.

Hardjoamidjojo, S. and R.W. Skaggs. 1982. "Predicting the Effects of Drainage System
Design on Corn Yields". Agricultural Water Management. 5:127-144.

Hardjoamidjojo, S. R.W. Skaggs, and GO. Schwab. 1982. "Corn Yield Response to

Excessive Soil Water Conditions". fIfranggions 91' the ASAE. St. Joseph,
Michigan: American Society of Agricultural Engineers. 25(4):922-927.

Harris, TR. and HF. Mapp. 1980. "A Control Theory Approach to Optimal Irrigation
Scheduling in the Oklahoma Panhandle." S ut e J on of icu tur

mm. 12:165-171.

Harris, T. R. and H. P. Mapp. 1986. "A Stochastic Dominance Comparison of Water-

Consomng Irrigation Strategiw" WWW
68(2): 298-305.

 

179

Harsh, S.B., L.J. Connor, and GD. Schwab. 1981. MW. Prentice-
Hall, Inc.: Englewood Cliffs, New Jersey.

He, C., T.C. Edens, B.E. Vieux. 1991. "Water Resource Allocation and Optimization for
Subsurface Irrigation in the Saginaw Valley, Michigan". Project Report Submitted
to the Soil Conservation Service. East Lansing, Michigan: Department of
Resource Development. Michigan State University.

He, C., T.E. Zahniser, B.K. Dillon, T.C. Edens. 1992. "Environmental Impacts of
Withdrawing Water from the Great Lakes for Agricultural Irrigation". Project
Report Submitted to the Soil Conservation Service. USDA. East Lansing,
Michigan: Institute of Water Research and the Department of Resource
Development. Michigan State University.

 

Huron County Extension Service. 1988 and 1990. "Huron County Integrated Farm
Management/ Crop Report". Cooperative Extension Service. Michigan State
University. East Lansing, MI.

Ignall, EJ. 1972. "On Experimental Designs for Computer Simulation Experiments."
Management Science. 18(7): 384-388.

Johnson, S. R. and G. C. Rausser. 1977. "Systems Analysis and Simulation: A Survey of
Applications 1n Agricultural and Resource Economics." A Survey 9f Agricultural
W. 2: 157-301. G. G. Judge et al., editors. Minneapolis:
University of Minnesota Press.

Just, R. E. 1974. "An Investigation of the Importance of Risk 1n Farmers’ Decisions."

AmsisnloemslsLAaisltusLEeonomLm-m

Just, R. E. 1975. "Risk Aversion under Profit Maximization." Man rignJ 011mg of
WW 57: 347-352

Just, RE. 1976. "Estimation of a Risk Response Model with Some Degree of Flexibility."

Wal- 42:675-684.
Just, RE. and RD. Pope. 1979. "Production Function Estimation and Related Risk
Considerations." er'can Jon 0 ic tural no 'cs. 61: 277-284.

King, RP. and LJ. Robison. 1981a. "Implementation of the Interval Approach to the
Measurement of Decision Maker Preferences." Michigan State University
Agricultural Experiment Station Research Report 418. East Lansing.

King, R. P. and LJ. Robison. 1981b. "An Interval Approach to Measuring Decision
Maker Preferencs"AmerisO_I.o.umsLoLAsO§elts1sEsmmis 63: 247-255

King, RP. and LJ. Robison. 1984. "Risk Efficiency Models." In: W
W. ed. PJ. Barry. Iowa State University Press.

180

Kittleson, K. and C. He. 1990. "Saginaw Bay Area Subirrigation/ Drainage Project". In
F. M. D’Itri and JA. Kubitz, eds. e Sa ' aw a ichi a

WWW. East Lansing: Institute of Water

Research.

Kittleson, K., C. He, C. Henshaw, and J. Ervin. 1990. "Subirrigation Regional Impacts:
Env1ronmental, Economic, Social". In F. M. D’Itri and J A Kubitz, eds. 1h;
11 - ~ " .East Lansing:

 

Institute of Water Research.

Kittleson, K., C. He, and J. lively. 1990. "Saginaw Valley Groundwater Resources:
Current Conditions and Trends". In F. M. D’Itri and J .A Kubitz, eds. 111;
- . ' 8. East Lansing:

 

Institute of Water Research.

Kramer, RA., and RD. Pope. 1981. "Participation in Farm Commodity Programs: A

Stochastic Dominance Analysis" WOW 63:
119-128.

Lazarus, W.F. and J.T. Scott, Jr. 1981. "A Present Value Analysis of Returns to Irrigation
in South Central Illinois." orth Central J on 0 ion tural onomics.
3(1):63-70.

LeCureux, J. P. and J. Booms. 1990a. "Plan of Work-- Project 111 Subirrigation System
Management". In F. M. D’ Itri and J A Kubitz, eds. The §aginaw Bay, Michigan

WW East Lansing Institute of Water

Research.

LeCureux, J. P. and K. Booms. 1990b. "Corn Production Management for Subirrigation".
In F. M. D’Itri and JA. Kubitz, eds. WW

8n hinigatinnzma ainage floject; 1987-88. East Lansing: Institute of Water
Research.

LeCureux, J. P. and K. Booms. 1990c. "Subirrigation Management for Sugar Beet
Production". In F. M. D’Itri and J A. Kubitz, eds. fihe 8ag'n1aw Bay, Michigan
WW East Lansing: Institute of Water

Research.

LeCureux, J. P. and J. Booms. 1990d. "Impact of Subirrigation on the Quality of
Discharge Water". In F. M. D’Itri and J A Kubitz, eds. jIhe Saginaw Bay,

WW 8East Lansing: Institute of

Water Research.

LeCureux, J. P. 1991a. "Profitability of Crop Rotation on a Sub1rr1gat1on/Dra1nage
System: 1989". In F. M. D’Itri and J A Kubitz, eds.

WW. East Lansing: Institute of Water

Research.

181

LeCureux, J. P. 1991b. "Profitability of Crap Rotation on a Sub1rr1gat1on/Dra1nage

System: 1990". In F. M. D’Itri and J A Kubitz, eds. mm
W. East Lansing: Institute of Water

Research.

Levins, RA. K.S. Wright, and L.G.Heatherly.1981.ASi1n111ation Approach to

W Department of Agricultural
Economics, Mississippi State University.

Liapis, PS. and LJ. Moffitt. 1983. "Economic Analysis of Cotton Integrated Pest

Management Strategies." Southern Jnurnal oi Aggicultural flanomics. 15(1): 97-
102.

Lynne, G.D., and R. Carriker. 1979. "Crop Response Information for Water Institutions."
J, 91 1h; [m'gatign and 12191.8. age Division Egoceedings of the ASCE. 105: 221-32.

Lynne, G.D., W.G. Boggess, and KM. Portier. 1984. "Irrigation Water Supply as a
Bioeconomic Process." §nuthem Jgumal 91 Agricultural Egnoniics. 16(2): 73-82.

Mapp, H.P., Jr., V.R. Eidman, J.R. Stone, and J.M. Davidson. 1975. Simulating Soil
W te t 0 he ' ess-Cro 'eld eatio s i s for Econo ic sis.
Oklahoma Agricultural Experiment Station Bulletin T140.

Meyer, J. 1977. "Choice Among Distributions." Log-11W. 14:326-336.

Michigan Department of Agriculture. 1958-1991. Agzigiggiitanm. Lansing,
Michigan.

Michigan Department of Agriculture. Environmental Division. Department of
Climatology. Maximum and Minimum Daily Temperature Data for Flint, MI:
1958 - 1990. East Lansing, Michigan.

Murphy, MC. 1968. "A Stochastic Approach to Investment Appraisal." Wig.
11(7):304-318.

Murugaboopathi, C., R.W. Skaggs, and RD. Evans. 1991. "Subirrigation for Corn and
Soybean Production for North Carolina Soils". Raleigh: Biological and
Agricultural Engineering Department. North Carolina State University.

Musser, W. N. and B. V. Tew 1984. "Use of Biophysical Simulation 1n Production
Boonontis" Wows 16(1) 77-36

Musser, W. N., B.V. Tew, and J ..E. Epperson. 1981. "An Economic Examination of an
Integrated Pest Management Production System with a Contrast between E-V
and Stochastic Dominance Analysis." 0 t J o 'c tur '
13:119-24.

 

182

Naylor. TH. 1971. W
New York: Wiley.

Nott, S. B., G. D. Schwab, A. E. Shapley, M. P. Kelsey, J. H. I-Iilker, L. O. Copland. 1992.
"1992 Crops and Livestock Budgets Estimates for Michigan" Agn'_nh_arai

W No. 556. Department of Agricultural Economics, Michigan
State University, E. Lansing, MI.

Pavelis, GA. 1987. "Economic Survey of Farm Drainage. In GA. Pavelis (ed. ) Eann
WW. USDA-ERS Miscellaneous Publication. 1455:110-

Phillips, B.H. 1969. "Rainfall Distributions and Simulation Models." 41st ANZAAS
Congress Section 8 Paper. Adelaide.

Pitt, RE. 1980. "Risk Assessment as a Decision-Making Tool for Agricultural Engineers."
American Society of Agricultural Engineers Paper Number NAR80-102.

Pope, RD. and RA. Kramer 1979. "Production Uncertainty and Factor Demands for the
Competitive Firm." Snuihenn mnomig Jnurnal. 46: 489-501.

Pope, R.D., and RF. Ziemer. 1984. "Stochastic Efficiency, Normality, and Sampling

Errors in Agricultural Risk Analysis." American J gonna] oi Agzicaitural
Egngmig. 66:31-40.

Pouliquen, LY. 1970. W. World Bank Staff Occasional
Papers Number 11. Baltimore: The Johns Hopkins Press.

Protasiewicz, L, H.W. Belcher, and GE. Merva. 1988. "Water Table Management Water
Quality Study." Final Report to the Michigan Department of Natural Resources.
Department of Agricultural Engineering, Michigan State University. East
Lansing, MI.

Protasiewicz, LJ. and David Auernhamer. 1991. "Water Table Management of Alfalfa".
In F M D’Itri and IA Kubitz, eds Wm
WW. East Lansing: Institute of Water

Research.

Raskin, R. and MJ. Cochran. 1986. "Interpretations and Transformations of Scale for the
Pratt -Arrow Absolute Risk Aversion Coefficient: Implications for Generalized

Stochastic Dominance" mMmfloLAawLowmows 11(2): 204-
210.

Reichelderfer, K.H. 1983. "Positive Farm Models: A Comment on Aesthetic Versus
Functional Value." WWW K H. Baum and
LP. Schertz, editors. Boulder: Westview Press. p. 281-285.

 

183

Robertson, J .D. W.N. Musser, and B.V. Tew. 1982. "Lease Versus Purchase of a Center
Pivot Irrigation System: A Georgie Example" WW
Egnnmig. 14(2): 37-42.

Rubinstein R.Y. 1981. Wag. New York, Chichester,
Brisbane, and Toronto: John Wiley & Sons.

Schoney, RA. and LR. Massie. 1981. "An Investment Feasibility Analysis of a Traveling

Gun Irrigation System 1n the Humid Midwest." WAD!
Wow 3(1) 53-61

Schreider, YA. 1966. W951. Oxford: Pergamon.

Selley, R. 1984. "Decision Rules in Risk Analysis." Chapter 5 Bisk Management n1'
Mingling. P.J. Barry, ed. Iowa State University Press. pp. 53-67.

Skaggs, R..W 1981. e re ce ds or v u t
Dginage-Waia; Ma anage man; 811s stains Ofor Sgiia wnh Mi gh Wate; fifabies. Reportf

prepared for the USDA Soil Conservation Service.

Skaggs, R.W., N.P. Fausey, and B.H. Nolte. 1981. "Water Management Model Evaluation
for North Central Ohio." Transactions of tha ASfl. 24(4):922-928.

Skaggs, R.W., S. Hardjoamidjojo, B.H. Wiser, EA. Hiler. 1982. "Simulation of Crop
Response to Surface and Subsurface Drainage Systems". ji'gnsagions 91 the
M. 25(6):1673-1678.

Skaggs, R.W. and A. Nassehzadeh-Tabrizi. 1983. "Optimum Drainage for Corn
Production". Technical Bulletin No. 274. Raleigh: North Carolina Agricultural
Research Service. North Carolina State University.

Skaggs, R.W. 1990. 123/ME MQD User’s Manual. North Carolina State University.
Raleigh, North Carolina.

Skaggs, R.W. 1991a. "Water Movement Factors Important to the Design and Operation
of Subirrigation Systems". WM. 24(6):]553-1561.

Skaggs, R. W. 1991b. "Modeling Water Table Response to Subirrigation and Drainage".
Irwsw'onuflhoASAE 34(1)=169-175

Smith, M. C., R. W. Skaggs, and J. E. Parsons. 1985." Subirrigation System Control For
Water Use Efficiency". W. 28(2): 489-496.

Spicer Engineering Company. 1990. "Water Supply for an Irrigation District: A
Feasibility Study". In F. M. D’Itri and J A Kubitz, eds. W

W East lensing Institute of

Water Research.

184

Taylor, C. R. 1986. "Risk Aversion versus Expected Profit Maximization with a

Progressive Income TaX" Amarianlanmalammaltaalmnamia 68(1) 137-
143.

Trebeck, D. B. and J. B. Hardaker. 1972. "The Integrated Use of Simulation and
Stochastic Programming for Whole Farm Planning Under Risk."A1_1§t_r_§1inn

IanmalaLAzriamtaraLEananna 16(2) 115 126

US. Department of Agriculture. Soil Conservation Service. 1982. 1982 Na_tnnnl
Baairsalnxemanz- Washington. DC.

US. Department of Agriculture. Soil Conservation Service. 1972. W
County, M1. Government Printing Office 0-388-653.

US. Department of Agriculture. Soil Conservation Service. 1980. §gil Suggy gt Hngon
W. Government Printing Office.

US. Department of Commerce. 1978 and 1987. Census n1 Agzigulture. Vol.1, part 22,
Michigan State and County Data. Bureau of the Census, US. Department of
Commerce, Washington, DC.

van Schilfgaarde, J. editor. 1974. W. Madison, Wisconsin:
American Society of Agronomists.

Whitmore, GA. 1970. ”Third—Degree Stochastic Dominance." m
Enanamiafiaiax 69(3)= 42-49-

Williams, Osminski, Little & Associates. 1990. "Huron County Subirrigation Water
Supply Feasibility Study - Phase II". In F.M. D’Itri and J A. Kubitz, eds. Ihg
Sngln' aw Bay, Michigan Subnz' igatiolemm’ nge E[niect: 1987-§8. East Lansing:

Institute of Water Research.

Williamson, K.D. & Associates, 1990. "Huron County Subirrigation Study: Feasibility

Report". In F. M. D’Itri and JA. Kubitz, eds. W
W. East Lansing: Institute of Water

Research.

Wilson, RN. and V.R. Eidman. 1983. '111e Financial Profitability of Irrigating Fine

Textured Soils in the Upper Corn Belt." 119W
W. 5(2):]03-110.

Worm, B. G., J. T. Ligon, T.V. Wilson, C. W. Doty, and E. E. Strickland. 1982. "Economic
Evaluation of Subsurface and Center Pivot Irrigation Systems." M
r ce v a Wate

ll 5 . 5. ”mm. s.

185

Yaron, D. and G. Strateener. 1973. "Wheat Response to Soil Moisture and the Optimal
Irrigation Policy Under Conditions of Unstable Rainfall." W
m. 9:1145-53.

Yassour, J. (1. Zilberman, and G. C. Rausser. 1981. "Optimal Choices Among Alternative

Technologies with Stochastic Yield" MW.
63:718-723.

Zacharias, T.P., and AH. Grube. 1984. "An Economic Evaluation of Weed Control
Methods Used in Combination with Crop Rotation: A Stochastic Dominance

Approach." W. 6:113-20-

Zusman, P. and A. Amiad. 1965. "Simulation: A Tool for Farm Planning Under
Conditions of Weather Uncertainty." MW. 47(3):547-549.

llliltllllllllllllli