124
969
THS

 

 

 

‘ - 1

i: .—---v. .. In.
Z"= .’.‘ r L:

l " ”"""' '
r- __

 

 

This is to certify that the
thesis entitled
The Conversion of Prime and GOod Farmland
To Developed Uses in Southern Michigan

presented by

Steven Carl Bennett

has been accepted towards fulfillment
of the requirements for

J ° 5 ' degree in End Use

Planning

 

Major professor

0.7639 MS U is an Ami-motive Action/Equal Opportunity Institution

III III

3 129300

II

 

MSU ‘
LIBRARIES
m

 

 

 

IIIII IIIIII’II’IIIIIW

 

 

 

 

638 9864

RETURNING MATERIALS:

_ Place in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped below.

 

7V

 

. ";“_I \.
. - E" “3 _ 9 ’g. .7 4
my??? 5 7935;

1,1. "*7
p.

its: 3;

I.r.\.‘.):£
a‘foy'
1:306 5; '
‘3' r — I“‘\
' rhl|I U U. l»‘»\)
f/

(War:

335”
—.
\0
so
\I

 

 

THE CONVERSION OF PRIME AND GOOD FARMLAND
TO DEVELOPED USES IN SOUTHERN MICHIGAN

By

Steven Carl Bennett

A THESIS

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

MASTER OF SCIENCE

Department of Crop and Soil Sciences

1986

ABSTRACT

THE CONVERSION OF PRIME AND GOOD FARMLAND TO
DEVELOPED USES IN SOUTHERN MICHIGAN

By

Steven Carl Bennett

Aerial photographs were used to determine the amount of prime and
good farmland converted to nonagricultural uses in southern Michigan
between 1962-1978. The study area consisted of the 41 southern-most
counties in lower Michigan. This area was divided into six regions based
on 1980 county populations and general soil characteristics. The
greatest amount of development in the study area occurred in Region 3
(containing Detroit, the nation's fifth largest metropolitan area) where
10 percent of the prime soils were developed between 1962-1978.
Development totals for the six regions ranged from 0.3 to 10.3 percent
on prime soils; 0.6 to 13 percent on good soils; and 0.9 to 6 percent on
"other" soils. Overall, 3 percent of the agricultural acreage in the

study area was deve10ped between 1962-78.

To my pal Buck, and Dr. John Shickluna.
May they rest in peace

11

ACKNOWLEDGMENTS

I would like to thank Rod Stocking fom the A808 office, Sherm
Hollander from the Michigan Department of Natural Resources, and the
Michigan State University Center for Remote Sensing Staff. Without
their cooperation and assistance this study would not have been
possible. I would also like to thank Dr. Milton Steinmueller and Dr.
Larry Libby for their assistance in the development of this thesis.

A special thank you to Dr. Jim Stapleton who helped develope the
Sampling techniques; to Dr. Jim Crum for all of his trouble-shooting
help and computer instruction; and to Wes Gaddis for his dedication to
the research. Wes made the days enjoyable and I am thankful for the
friendship that we developed. A

I would like to thank my father, Carl Bennett, for his help with
statistical problems and assistance in generating many of the ideas used
in this study. More importantly, he was the example of a graduate
student that I tried to follow.

I can not thank Dr. Del Mokma enough for all of his help and
guidance. I learned a great deal from him and he forced me to learn a
great deal on my own. Del treated all of his students as equals and
provided an example for other major professors to follow.

I am very grateful to my wife, Kristy, who endured both her
schooling and mine. Kristy provided a tremendous amount of support,
help, and advice; if there was a problem I could always count on her for

the right answer.

iii

TABLE OF CONTENTS

Page

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . v

LIST OF FIGURES O O O O O O O O O O O O O O I O O O O O O O O O O 0 Vi

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Land Use Information . . . . . . . . . . . . . . . . . . . . . 4
Sources and Problems . . . . . . . . . . . . . . . . . . . 4
Land Use Data . . . . . . . . . . . . . . . . . . . . . . 11
Michigan Data . . . . . . . . . . . . ...... . . . . l7
Causes of Rural Land Conversion . . . . . . . . . ...... . 19
Interpreting the Data . . . . . . . . . . . . . . . . . . . . . 21
Export Projections . . . . . . . . . . . . . . . . . . . . 21
Yield Trends and Projections . . . . . . . . . . . . . . 22
Other Demands for Agricultural Land . . . . . . . . . . . 24
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 25

STUDY AREA AND METHODS . . . . . . . . . . . . . . . . . . . . . . . 29

RESULTS AND DISCUSSION . . . . . . . . . . . . . . . . . . .'. . . . 37

CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

FUTURE RESEARCH . . . . . . . . . . . . . . . . . . . . . . . . 48

LIST OF REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . 49

APPENDIX A: COUNTY DATA SOURCES . . . . . . . . . . . . . . . . . . 53

APPENDIX B: SOIL ACREAGES BY REGION AND COUNTY . . . . . . . . . . 55

iv

Table

10

11

LIST OF TABLES

Page
Agricultural Land Conversion Rates . . . . . . . . . . . . . 12
National Land Use Data . . . . . . . . . . . . . . . . . . . 14
Michigan Land Use Data . . . . ..... . . . . . . . . . . 19
Total Conversion of Agricultural Land Between 1962-1978 . . 37
Farmland Conversion by Population Distribution . . . . . . . 38
Total Land Conversion by Soils between 1962-1978 . . . . . . 39
Percentage of Development by Soils . . . . . . . ...... 40
Distribution of Prime, Good, and Other soils in Each Region . 40
Percentage of Prime, Good, and Other Soils Developed . . . . 41
County Data Sources . . . . . . . . . . . . . . . . . . . . . ‘53

Soil Acreages by Region and County . . . . . . . . . . . . . 49

Figure

LIST OF FIGURES

Study Area . . . . . . . . . . . . . . . . . . . . . .
Study Area with Cities . . . . . .

Farmland Conversion in Huron County

Farmland Conversion in Berrien County . . . . . . . . . .
Farmland Conversion in Bay County . . . . . . . . . . . . .

Farmland Conversion Rates in Huron, Berrien and Bay Counties

vi

Page
30
31
42
43

44

INTRODUCTION

Prime Farmland is land that has the best combination of physical
and chemical characteristics for producing food, feed, fiber, forage,
and oilseed crops (Hilner, 1985). It can be cropland, pastureland,
rangeland, or forest land, but not built-up land or in a water use. It
has the soil quality, growing season, and moisture supply needed to
economically produce sustained high yields of crops using acceptable
farming practices. Prime farmland has an adequate and dependable water
supply from precipitation or irrigation, a favorable temperature and
growing season, an acceptable pH, acceptable salt and sodium levels, few
or no rocks, and good permeability to water and air. Prime farmland is
not excessively erodible or saturated with water for extended periods
of time. It does not flood frequently and is not protected from
flooding.

Good farmland in Michigan is land that has physical and chemical
properties capable of producing good yields of food, feed, fiber,
forage, and oilseed crops. It has steeper slopes, greater wetness
problems, lower available water, and higher production costs than prime
farmland (Hilner, 1983). Prime and good farmlands are considered
important for crop production at the national and local levels.

In 1976 the Soil Conservation Service (808) reported that 3 million

acres of agricultural land were being converted annually to

2

nonagricultural uses, an 82 percent increase from the previous decade
(Brewer and Boxley, 1982). The annual cropland conversion was 675,000
acres, of which 32 percent was prime farmland (Sampson, 1981; Brewer and
Boxley, 1982). The nonagricultural uses included roads, houses,
businesses, airports, and water projects. Brewer and Boxley (1982)
considered these uses irreversible. The SCS report, coupled with a
dramatic increase in foreign export demand and the 1973 OPEC oil
embargo, raised doubts about the ability of the United States to meet
future demands for food and fiber (Huffman, 1981; Brewer and Boxley,
1982; Crosson, 1982a; Fischel, 1982). These uncertainties prompted the
establishment of the National Agricultural Lands Study (NALS).

The National Agricultural Lands Study was conducted by the U.S.
Department of Agriculture (USDA) and the Presidents Council on
Environmental Quality (CEO). The objectives were 1) to study the
availability of the nation's agricultural lands 2) determine the nature,
rate, extent, and causes of agricultural land conversion and 3) evaluate
the economic, environmental, and social costs of agricultural land
conversion (Sampson, 1981; NALS, 1981). The NALS used data from the
SCS, the Census of Agriculture, the U.S. Forest Service, and the
Economic and Statistics Service (ESS) to identify changes in major land
uses (NALS, 1981).

The Soil Conservation Service collected land use information in the
1958 and 1967 Conservation Needs Inventories (CNI), the 1975 Potential
Croplands Study (PCS), and the 1977 and 1982 National Resource
Inventories (NRI) (McGill et a1., 1981). McGill et a1. (1981), Brewer

and Boxley (1982), Fischel (1982), Raup (1982), and Frey (Easterbrook,

3

1986) identified a number of problems with the SCS studies and the NALS.
Brewer and Boxley (1982) noted that the individual SCS studies were not
comparable or comprehensive because they were tailored for specific SCS
programs, the procedures were different, they were conducted at
different levels of statistical and geographical reliability (McGill et
a1., 1981), and the perceptions of land use changed after the 1958 and
1967 CNI's (Raup, 1982; Easterbrook, 1986). Because of these
differences, Fischel (1982) believed that the NALS (which relied heavily
on SCS data) overstated the conversion of rural land to urban uses by a
factor of 2 and probably 3 or 4.

The 1982 NRI (Easterbrook, 1986) concluded that earlier estimates
of farmland conversion were "markedly overstated." Proponents of
farmland preservation (Misseldine and Riggle, 1985; Reganold, 1986)
however, continued to cite the 3 million acre per year figure
(Easterbrook, 1986) reported by the SCS and the NALS.

The data base on agricultural land conversion is questionable
(Brewer and Boxley, 1982; Fischel, 1982; Easterbrook, 1986). Methods
used to collect the data contain many errors and the data reported by
different agencies are not comparable. Proponents of farmland
preservation, however, frequently cite the inadequate data, hoping to
have farmland preservation policies adopted. The data base needs to be
updated using highly objective methods designed to measure agricultural
land conversion before any policy decisions can be made. The most
efficient technique for gathering accurate land use data is by using
aerial photography (Fischel, 1982; Easterbrook, 1986). The objectives

of this study were 1) to develop an inexpensive, objective technique

E#— ”.371le

4
using aerial photography for measuring the conversion of
agricultural land to developed uses, 2) to use this technique to measure
the conversion of agricultural land to developed uses in southern
Michigan between 19621 and 1978, and 3) to determine farmland conversion

rates over a 30 year period in three representative counties.

LAND USE INFORMATION

Sources and Problems

Urban and Built Up Land was defined in the National Agricultural
Lands Study (1981) as land used for residences, industrial sites,
commercial sites, construction sites, railroad yards, parks of fewer
than ten acres within urban areas, cemeteries, airports, golf courses,
sanitary landfills, sewage treatment plants, water control structures
and spillways, reservoirs, highways and roads, shooting ranges, and
similar areas. This operational definition is used throughout this
review.

The Soil Conservation Service provided land use information in the
1958 and 1967 Conservation Needs Inventories (CNI), the 1975 Potential
Croplands Study (PCS), and the 1977 and 1982 National Resource
Inventories (NRI). These studies have frequently been used to quantify
land use changes over time (McGill et al., 1981).

The 1958 CNI was a two percent sample of nonfederal rural lands.
The princicpal objective was to identify areas needing conservation
treatment. The study examined 170,000 randomly selected primary

1
Dates varied between 1958-64 (see Appendix A)

5
sampling units (PSU's) in the field. The average PSU was 160 acres.
The data generated were statistically reliable at the county, state,
regional, and national levels. The study excluded from field
examination: urban and built up areas of more than ten acres, federal
land (except cropland), and bodies of water (McGill et al., 1981;
Brewer and Boxley, 1982; Raup, 1982). County records describing
incorporated areas, maps, and U.S. Census data were used to obtain and
check estimates of urbanized acreages in selected counties (Brewer and
Boxley, 1982; Fischel, 1982).

The 1967 CNI was an update of the 1958 estimates. The principal
objectives and sampling frame remained the same. The sampling rates,
varied between one and eight percent nationally (McGill et al., 1981).

The use of county records (describing incorporated areas) as an
enumeration technique may have caused errors in the CNI's (Brewer and
Boxley, 1982). This method may have lead to overestimation of urban and
built up areas since all land in the incorporated boundaries would be
classified as developed, even if they were not. If this procedure were
used for an entire state, however, unincorporated - built up areas would
not have been included and, therefore, urban and built up areas would
have been underestimated. The extent to which these procedures were
used in the 1958 and 1967 CNI's is not known, thus both studies contain
unspecified errors.

The 1975 Potential Cropland Study (PCS) was a 0.05 percent sample of
the nation's nonfederal land. The principal objective was to quantify
the potential for new cropland and to identify problems with bringing

new land into production. A subsample of data points from the 1967 CNI

6
was used. Five hundred-six counties were randomly selected nationwide,
with each state having at least one county selected. There were less
than 5,000 PSU's. District conservationists (who conducted the study)
paid particular attention to land use changes since 1967, including
changes to urban and built up uses (McGill et al., 1981; Brewer and
Boxley, 1982; Fischel, 1982; Raup, 1982).

Some assumptions changed between the 1967 CNI and 1975 PCS
(Easterbrook, 1986). In 1967 many 10 to 40 acre urban areas were
classified as "other rural land", and then in 1975 they were
reclassified as urban. This would result in an overestimate of
urbanized areas (Fischel, 1982; Easterbrook, 1986). McGill et al.
(1981) and Raup (1982) believed that the problem was most pronounced at
the urban-rural fringe where there was scattered-lot, low density
develOpment. Fischel (1982) also noted that census data were not used
to check urban area estimates, as was the case in the 1967 CNI.

The sample size in the PCS was too small and the statistical
confidence limits too narrow for detecting urban uses and then expanding
them to regional and national levels (Raup, 1982; Fischel, 1982). There
were 41,000 sample points from 1967 revisited, but they were selected in
such a way that there were nine sample points to each plot of land. The
plots averaged 160 acres. The urban land recorded was 1.17 percent of
the total area examined. This is 480 (.0117 x 41,000) urban or built up
points. But most urban development typically takes up an entire plot
(all nine points) so the true number of urban observations was closer to

1/9 of 480, or 53 (Fischel, 1982).

7
The 1977 National Resources Inventory was a one percent sample
of all nonfederal land - both urban and rural. A wide range of concerns
were addressed, including wetland availability, soil erosion, and
potential wildlife habitat. The principal objective was to obtain data
for long - range soil and water conservation planning and programming.
There were 70,000 PSU's (ranging from 40-640 acres) randomly selected
from every county in 48 states. The average PSU was 160 acres. Three
points were selected as subsamples within each PSU (only two points were
used in PSU's of 40 acres). A three percent subsample, or 210,000
points, was selected for on-site investigations. The investigations
identified physical characteristics and changes in land and water area
since 1970. The data were statistically reliable at the state, farm
production region, and national levels (McGill et al., 1981; Sampson,
1981; Brewer and Boxley, 1982). Urban, built up, and incorporated areas
of more than 40 acres were derived from recent USGS, highway, or
planning commission maps of the respective counties. Aerial estimates
of these uses of less than 40 acres were obtained from the PSU's
(Sampson, 1981; Brewer and Boxley, 1982; Fischel, 1982). Bureau of the
Census data were not used because they often excluded developments of
less than 40 acres. The minimum population for a census Urban Place is
2500, and the area is seldom less than 600 acres (Fischel, 1982).
Fischel (1982) believed that the 1977 NRI procedures were biased
and allowed overestimates of the rate of urbanization when compared to
1967 CNI procedures. (It should be noted, however, that the NRI did not
calculate changes in agricultural land use since 1967; this was done in

the National Agricultural Lands Study, which is discussed below). In

8

1967 the District Conservationists referred to U.S. Census data to check
estimates of urban areas. Also, in 1967 all urban acreages were
obtained from USGS, highway, or planning commission maps. In the NRI
only urban areas of greater than 40 acres were measured from maps. A
final source of overestimation occured where farmland was adjacent to
urban areas (Fischel, 1982). In 1967 this land was excluded from the
urban total. In 1977, this land was included in determining urban
areas. .

The 1982 National Resources Inventory was the most extensive study
of natural resources on non-federal lands ever conducted by the SCS. It
contained facts on 22 natural resource components, including wind and
water erosion, the amount of prime farmland, the amount of potential
cropland, and land use. There were 350,000 PSU's nationally. The
average PSU was 160 acres with three subsamples taken from each. The
information was statistically reliable down to the substate (multi-

county) level (U.S.D.A., 1984).

\

Quantifying land use changes over time requires comparable measures
of use at two points. The SCS studies are commonly used for this even
though the data were not always comparable or .comprehensive. The
individual studies were tailored for specific SCS programs (Brewer and
Boxley, 1982), the procedures were different (Brewer and Boxley, 1982;
Fischel, 1982; Fischel, 1982), the studies were conducted at different
levels of statistical and geographical reliability (McGill et al.,
1981), and the perceptions of land use were different after the 1958 and
1967 CNI's (Raup, 1982). Fischel (1982) believed that measuring

increments between the 1967 CNI and the 1975 and 1977 studies was

I
I

9

inappropriate because of these differences, and large overestimates
occurred. SCS personnel agreed that the data collecting procedures
were different, but they did not believe that large overestimates
resulted. McGill et al. (1981) noted that all data were collected
under conditions of limited funding and personnel, but within those
constraints SCS provided the most reliable sampling frame possible.

The National Agricultural Lands Study was conducted in 1979 by the
U.S. Department of Agriculture (USDA) and the Presidents Council on
Environmental Quality (CEO). The objectives were: 1) to study the
availability of the nation's agricultural lands; 2) determine the
,nature, rate, extent, and causes of agricultural land conversion; 3)
evaluate-the economic, environmental, and social costs of agricultural
land conversion; and 4) recommend administrative and legislative
actions, if necessary, to reduCe the problem (Sampson, 1981; NALS,
1981). The NALS used data from different federal agencies to identify
changes in major land uses. The sources included SCS, the Census of
Agriculture, the U.S. Forest Service, and the Economic and Statistics
Service (ESS) (NALS, 1981). Fischel (1982), however, believed the NALS
relied exclusively on SCS data for reporting the conversion of rural
land and the size of the agricultural land base.

The information in the NALS was fragmentary and, in some cases,
inconsistent because the purposes, methods, and timespans of the data
differed (McGill et al., 1981; Brewer, 1981). Fischel (1982) believed
that the differences in procedures and objectives between the SCS
studies should have precluded their heavy use in the NALS. He

contended that the NALS overstated the conversion of rural land to urban

10

uses by at least a factor of two and probably three or four. Fischel
(1982) also believed that much of what the NALS called rural population
increase was most likely located in high density suburban areas and
unincorporated towns, which are truly urban even though unincorporated.
Despite these shortcomings, the NALS was indicative of the land use
problems in Illinois (Dovring et al., 1982), and Wheeler and Harper
(1983) believed that no one had offered any data more reliable than the
NALS estimates of farmland conversion.

The Economic and Statistics Service (ESS) assembled land use data in
its Major Land Use Series (1949, 1954, 1959, 1964, 1969, 1974). The
ESS did not generate its own data, however. It relied on data from
agencies such as the Soil Conservation Service, the Bureau of the
Census, the Forest Service, U.S. Geological Survey, the Bureau of Land
Management, the Census of Agriculture, and the Crop Reporting Board
(which is ‘in the ESS) (McGill et al., 1981). The classification
intervals coincide with the Census of Agriculture. The primary
objectives were to interpret and reconcile current data generated by
different agencies, relate the new estimates to earlier estimates, and
produce a time series.

Problems with the E58 data include: 1) the acreage estimates were
derived from reports made by the respective agencies and thus reflected
the agencies‘ determination of land use; 2) the study was primarily an
economic survey, not a physical one; 3) because Bureau of the Census
data were used, communities with a population under 2500 outside of
urbanized areas were not included (Boxley, 1981); 4) since estimates

were based on a variety of sources, an overall statement of

11

statistical reliabilty was not possible (McGill et al., 1981); and 5)
inconsistencies in land use definitions occurred (McGill et al., 1981).

The Census of Agriculture was taken at five year intervals (1949,
1954, 1959, 1964, 1969, 1974, 1978) to provide a continual source of
data on farmland characteristics (McGill et al., 1981). Some of the
characteristics included the number of farms, total cropland, and total
cropland harvested. These characteristics provided economic and social
information related to agricultural activities. Until 1969, data were
obtained by personal interviews of land users. The current procedure is
to send questionnaires to occupiers of land (McGill et al., 1981;
Dovring et al., 1982).

The problem with the Census of Agriculture is that total land
area is derived from deeds and leases (Dovring et al., 1982). The
information in deeds and leases was based on measurements made at the
time of the original rectangular survey. In Illinois this occurred
between 1813 and 1849. These measurements contained many errors and
were placed in platbooks. In Illinois the 1978 Census overcounted

farmland by five percent in about two - thirds of the counties.

Land Use Data

The Potential Cropland Study determined that between 1967 and 1975
23.3 million acres of agricultural land were converted to
nonagricultural uses, with an annual conversion rate of 3 million acres
(NALS, 1981; Sampson, 1981; Brewer and Boxley, 1982) (Table 1).
Approximately 16.7 million acres (70 Z) were converted to urban and
built-up uses, and 6.7 million acres (30 2) were converted to water

projects. The total cropland loss was 5.4 million acres or 675,000

12
acres per year (Table 1). Sixty percent of the cropland was in SCS
Classes I - III and 32 percent was prime farmland. During the same
period, 48.7 million acres were converted to cropland- These figures
represent an 82 percent increase in the conversion rate of agricultural

land compared to the period 1958-67 (Brewer and Boxley, 1982).

Table 1 Agricultural Land Conversion Rates

TOTAL ANNUAL TOTAL ANNUAL HARVESTED ANNUAL
SOURCE DATES AG.LAND AG.LAND CROPLAND CROPLAND CROPLAND FARMLAND

---------------------- Million acres---------------------

1
SCS 1958-67 1.9
1967-75 23.3 3.0 5.4 675,000 (thousand)
1967-77 25.0 2.5
2
NALS 1915-55 1 345
1956-72 289
1972-80 353
3
CENS. 1974-78 +18.2 +4.55 - 8.65
OF 1969-74 -17.1 -3.42 -10.70
AG. 1969-78 + 1.1 +0.12 - 9.79

1 NALS, 1981; Sampson, 1981; Brewer and Boxley, 1982; U.S.D.A., 1982.
2 NALS, 1981.
3 McGill et al., 1981; Healy, 1982.

Publication of the SCS data was not governed by any established
protocol to assure a clear statement of the results and their
limitations (Brewer, 1981; Caughlin, 1981; McGill et al., 1981)

According to Clark (1979), the United States lost 35,000 acres of its

agricultural land base each week, land sufficient enough to feed 100,000

13

Latin Americans, Africans, or Asians for a year. Jeffords (1979) stated
that if the decline in agricultural land continued, within 10 or 20
years the U.S. would lose its ability to feed the world and even its own
citizens. Most of these statements assumed that 3 million acres of
cropland were being paved over each year (Raup, 1982), not 675,000
acres, which was only 0.17 percent of the total cropland reported in
the 1977 NRI (Brewer and Boxley, 1982).

The NALS reported an annual cropland conversion rate of
675,000 acres (Table 1) from a cropland base of 413 million acres (Table
2). The study identified approximately 500 million acres of federally
owned agricultural land containing about 466,000 acres of cropland, and
1.36 billion acres of nonfederally owned agricultural land (NALS, 1981).
One hundred twenty - seven million acres were identified as potential
cropland. Thirty-six million acres had "high" potential and 91 million
acres had "medium potential". Land with "high potential" had
physical characteristics conducive to production, and similar lands in
the area must have been brought into production within three years prior
to 1975. Land with "medium potential" had favorable physical

characteristics but higher erosion potential and conversion costs.

14

Table 2 National Land Use Data
Harvested Total Urban & Built-up
Source Dates CrOpland Cropland & Transportation

---------------- Million acres---------------
1

SCS 1958 448 51
1967 438 61
1975 400 78
1977 413 90
1982 421 47
2
NALS 1955 345
1972 289
1977 413
1980 353 65
3
E88 1949 352 409» 38
1954 338 -
1959 317 399 50
1964 291 -
1969 286 384 60
1974 322 382 64
4
CENSUS 1949 345 478
OF AG. 1954 333 460
1959 311 448
1964 287 434
1969 273 459
1974 303 440
1978 462
5
CROP 1949 352
REPORT- 1954 338
INC 1959 317
BOARD 1964 291
1969 284
1974 320

1 McGill et al., 1981; Brewer and Boxley, 1982; U.S.D.A., 1982;
Easterbrook, 1986.

NALS, 1981; Easterbrook, 1986.

McGill et al., 1981.

McGill et al., 1981; NALS, 1981.

McGill et al., 1981.’ -

U'lbUJN

15

The NALS (1981) reported that acreages of harvested cropland
remained stable between 1915 and 1955 at approximately 345 million
acres. In 1956 these acreages began a seventeen year decline, falling
to 289 million in 1972. Between 1972 and 1980 foreign export demand
increased, and harvested cropland increased to 353 million acres (Table
2). These results, however, are not consistent with other data.

The Soil Conservation Service reported cropland acreage steadily
decreasing from 448 million acres in the 1958 CNI to 400 million acres
in the 1975 PCS (Table 2), with an annual conversion rate of 2.8 million
acres (Table 1). Between 1975 and the 1977 NRI, the SCS showed cropland
.increasing to 413 million acres (Table 2) (McGill et al., 1981).

The- Economic and Statistics Service reported a sharp decline in
cropland from 478 million acres in 1949 to 465 million in 1974, with an
increase in the late 1970's (Table 2) (McGill et al., 1981). Eighty
percent of the decline occurred between 1949 and 1964.

The U.S. Bureau of the Census reported that from 1969 to 1974 the
amount of land converted to nonagricultural uses grew at almost exactly
the same yearly rate as it did from 1959 to 1969 (Table 2) (Healy,
1982). The annual conversion rate was about 1 million acres (Table 1)
(McGill et al., 1981).

The Census of Agriculture indicated that between 1969 and 1978 over
35 million acres of cropland shifted out of and back into cropland, with
about a one million acre net gain (Table 1) (NALS, 1981). Census of
Agriculture data are often cited to dispute the NALS (Dovring et al.,

1982).

16

The 1978 Census of Agriculture has 1.05 billion acres of "land in
farms" with 460 million acres of cropland (Table 2) (NALS, 1981). The
ESS reports 464 million acres of cultivated cropland in 1974 with 37
million acres of urban and built-up land (Table 2). The 1977 NRI
indicates 413 million acres of cultivated cropland with 63 million acres
of urban and built-up land (Table 2). These discrepancies were explained
by differences in the definition and classification of data elements,
differences in the geographic areas and land uses inventoried, and
differences in the time periods covered (Brewer, 1981; McGill et al.,
1981; NALS, 1981).

The 1982 NRI concluded that earlier estimates of farmland
development were "markedly overstated." It reported an urban land total
of 47 million acres and a continuation of the 1967 to 1977 cropland
increase of 395 million and 413 million acres, respectively to 421
million acres (Table 2) (Easterbrook, 1986).

Zeimetz et al. (1976) conducted a study of the 53 fastest growing
counties in the U.S. to identify land use changes. The study compared
Agricultural Stabilization and Conservation Service (ASCS) 1:20,000
scale aerial photographs from 1961 and 1970. The selected counties
accounted for 20 percent of the 1960 to 1970 population increase. The
study found urban development occupied 16.4 percent of the land area in
1970, a 3 percent increase from 1961. Of the developed land, 35
percent was cropland; however, there were regional variations between 6
and 70 percent nationwide. The ten year increase in urbanization was 23
percent (3/13). The NALS estimated a 47 percent increase in urban

area for the entire U.S. from 1967 to 1977. If all of the U.S.

17
population increases in the 1960's occurred at identical rates as the 53
counties in the Zeimetz et al. study, 427,800 acres would have been
converted each year (Fischel, 1982). The NALS was aware of the Zeimetz
et al. study, and Fischel wondered why the NALS failed to spend part of
its 2.2 million dollar budget on updating the data.

The U.S. Geological Survey (USGS) is classifying the land area of
the U.S. using high altitude (not satelite) photo mapping (McGill et
al., 1981; Fischel, 1982). The study began in 1975, and in August 1980,
four states were completed. County development patterns and state totals
for urban areas of greater than 10 acres were compared with raw data
from the 1977 NRI. The comparison indicated that the NRI estimates
exceeded the USGS data by 29 percent in Pennsylvania, 41 percent in West
Virginia, and 74 percent in Florida. In Kansas the NRI data were
approximately the same as the USGS estimates. However, Kansas was
classified by the SCS as having negative conversion of agricultural land
between 1967 and 1977 (Fischel, 1982). The major discrepancies occurred
in a few counties. Urban development, however, was concentrated in only
a few counties. Therefore, a few overestimates in these counties

could cause a significant overestimate for the state.

Michigan Data

Large regional variations in the urbanization of agricultural land
were reported in the NALS Final Report (1981). Brewer and Boxley (1982)
stated that the problems associated with cropland conversion occurred
locally and regionally as well as nationally, and local and regional
data were needed. They believed that national figures tended to maskxf

more extreme situations at local and regional levels.

18

Land use data for Michigan were available from the 1958 and 1967
SCS Conservation Needs Inventories, the 1977 and 1982 SCS National
Resources Inventories, the Census of Agriculture, and the Michigan
Department of Agriculture (MDA). The MDA assembled land use data
annually in its Michigan Agricultural Statistics which provide economic
and production statistics for Michigan agricultural products. (A summary
of the data from these sources is found in Table 3. Misseldine and
Riggle (1985) reported farmland conversion using Census of Agriculture

and 1982 NRI data. They documented an annual prime farmland conversionxx

rate of 120,000 acres between 1977 and 1982 (Table 3).

19

Table 3 Michigan Land Use Data

Land in Urban Total
Source Dates Farms Areas Cropland
------------- Million Acres------------
1
Mich. 1958 16.0
Ag. 1959 819 (thousand) 10.8
Stats. 1960 15.4
1965 14.1
1967 13.6
1975 11.8
1977 11.6
1978 11.4
' 2
.Census 1935 18.5
of Ag. 1945 18.4
1950 17.3 11.0
1959 14.8 10.0
1964 13.6 9.5
1969 12.0 8.6
1978 11.5 8.4
1982 10.9 8.5
3
SCS 1958 2.0 10.5
1967 2.6 11.3
1977 3.8 9.5
1982 2.0 9.4

1 Mich. Dept. of Ag., 1959; Wright, 1974; M.D.A., 1976; M.D.A., 1978.

2 Wright and Caul, 1967; Wright, 1974; Wright, 1984; Misseldine and
Riggle, 1985.

3 U.S.D.A., 1968; U.S.D.A., 1982; U.S.D.A., 1984.

\,a
3..

CAUSES OF RURAL LAND CONVERSION I

\:

V-~.‘.

Population demographics contributed significantly to the conversion:::k
of rural lands (Brown and Beale, 1981; Brewer and Boxley, 1982). Between
1970-79, the population around urban centers grew at an average
annual rate of 1.6 percent, while the annual population growth rate of
inner cities was less than one percent. Between 1970 and 1978

nonmetropolitan areas averaged 10.2 percent growth compared to 6.1

20
percent growth for metropolitan areas (Sampson, 1981). Rural areas also
experienced a significant growth rate, averaging 9.6 percent. From 1970
to 1976 2.3 million more people moved to nonmetropolitan counties than
moved out of them. This trend was reversed from the 1960's when 3
million people moved out of nonmetropolitan areas (Brewer and Boxley,
1982).

Forty - three percent of the housing starts from 1970 to 1977 were
in rural areas, and two - thirds were located on lots with septic
systems. Lots with septic systems typically require more land than lots
with central sewer systems (Brewer, 1981; NALS, 1981; Brewer and Boxley,
1982). The average size of a sewered lot was roughly 0.25 acres; the
average size of an unsewered lot was roughly 2 acres (Brown and Beale,
1981).

There was a 22 percent increase in the number of households between
1970 and 1979 (NALS, 1981) and an increase in scattered - site rural
development (Brown and Beale, 1981). Fischel (1982) did not believe
that scattered-site rural development was a significant cause of rural
land conversion. He stated "can one seriously argue that a single -
family home on a multi-acre lot causes the entire area to be 'paved
over or built up?‘" Fischel (1982) failed to consider that efficient
agricultural production requires large land units (Whiteside and
Schaner, 1972).

Highways covered approximately 21.4 million acres of rural land in
1977, an increase of 458,000 acres from 1964 (Brewer and Boxley, 1982).
Since most growth was prior to 1970, Brewer and Boxley (1982) did not

believe that highway construction was a significant cause of the

21
increased rate of rural land conversion reported by the NALS.

In 1979, about 1.5 million acres of land were disturbed by the
surface mining of coal (Brown and Beale, 1981). Brewer and Boxley
(1982), however, did not believe that surface mining of coal, or rights
of way for energy use significantly contributed to the increased

conversion of rural land.

INTERPRETING THE DATA

The problem with land use data is that the net effects of””*

agricultural land conversion on total agricultural production are
overlooked (Sampson, 1981). The data must be interpreted asia factor of
1) growth in demand for food, fiber, and land, 2) increases in crop
yields, '3) quality and quantity of land in reserve, and 4) the cost of
bringing that land into production (NALS, 1981; Sampson, 1981; Crosson,

1982b). There are many interpretations of these 4 factors.

Export Projections
Exports are the driving force behind expanding agricultural 7\
production and seem to determine the future demands for agricultural
land (Huffman 1981). The NALS Final Report (1981) predicted the large
agricultural export demand experienced in the 1970's to increase to 2000
and place most of the nation's 540 million acres of cropland into
cultivation. These projections were made using the USDA - CARD model.
In terms of volume, the NALS predicted that export demand could triple
between 1980 and 2000. With little change in total cropland acreage,
however, production in 1981 reached 92.5 percent of the NALS volume

projections for wheat and 94.2 percent for feed grains. And in

22
1981/82 the percentages of production used for exports exceeded the
2000 projections for wheat and feed grains (Raup, 1982).

Export demand is dependent upon developments in exporting countries.
Some countries, including Russia and the Peoples Republic of China, want
to increase domestic food production and rely on imports only when
domestic production is low (Crosson, 1982a). Heady (1982) reported
projections from the National Research Council indicating that food
production could double in developing countries by 2000; and in high
income countries, grain production could be increased beyond current
U.S. production, thus reducing foriegn demands upon the U.S.

agricultural land base.

Yield Trends And Projections

Crop yield trends are viewed in many different ways. Some people
believe yield growth curves are beginnig to flatten; others believe the
curves are still increasing but at a rate paralleling the 1935 to 1955
period. Yield trends depend upon the time span viewed. In general, the
shorter the timespan the more pessimistic the view (Heady, 1982). Yield
trends based on observations made up to the mid - 1970's were the most
pessimistic; observations made through 1981 were more optimistic.
Healy (1982) saw no convincing evidence indicating yields have plateaued
because yields and total productivity in the 1970's were drastically
affected by fluctations in the weather. Crosson (1982b) indicated yield
trends since 1972 decreased relative to the pre - 1972 period, but this
did not mean that yield increasing technologies or production
potentials were being achieved. Yields did not increase as rapidly

partly because high energy technologies developed after 1972 were too

23
expensive to use. Crosson (1982b) also noted that the increase in
cropland acreage after 1972 was on poorer quality- soils, so the
flattening of the curves could be due to farmers switching from
intensive to extensive management practices. Heady (1982) believed the
differences betweeen average yields and record yields indicate that
production potentials have not been achieved.

Crosson (1982a) pedicted yields of grain and soybeans to increase
1.2 percent from-1979 to 2005. He also believed that if yields continue
to grow at the same rates, projected output could be met without
increases in cropland acreage. Abel (1982), however did not expect any
major technological breakthroughs in crop production in the near future;
therefore, yields would not increase with rising export demand and more
acreage would be needed (NALS, 1981). Boxley (1981) noted the U.S. will
have adequate land resources to meet domestic needs for food, fiber,
timber, and housing but foriegn demand must be considered.

There is uncertainty about the role of technology in increasing.
yields because of rising real prices for fertilizers and energy, and
lack of unutilized water supplies for irrigation (NALS, 1981). The use
of irrigation peaked in the 1970's and further - large productivity
gains were not to be expected (Heady, 1982). Pump - irrigation use is
slowing and is expected to decrease across the Great Plains as energy
prices increase and water supplies decrease (Heady, 1982). The High
Plains Study (Supalla et al., 1982), however, estimated that irrigated
acreage in the Ogallala region will increase to 2000 and production of
all major crops will increase by at least 50 percent. The study

reported technology will increase crop yields and irrigation efficiency-

24
allowing productivity increases at lower costs, despite declines in
available water sources.

Real prices of energy and fertilizer are expected to increase
production costs (Crosson, 1982b). The High Plains Study (Supalla et
al., 1982) projected energy costs to increase from 1977 to 2020 by 137
percent for diesel fuel, 309 percent for natural gas, 112 percent for
propane, and 94 percent for electricity. The cost of nitrogen is
expected to increase by 66 percent of the increase in natural gas
prices. Herbicide costs are expected to decrease by ten percent (in
real terms) through 1990 and remain constant thereafter (Supalla et al.,
1982). According to Heady (1982), agricultural demand for energy is
very inelastic, except for pump - irrigated crops. So, in general,

rising energy prices do not effect yields directly.

Other Demands For Agricultural Land

If energy prices skyrocket, the use of agricultural land for energylz
production could be extensive, such as the use of corn for ethanol
production (Heady, 1982; Abel, 1982). The NALS Final Report (1981)
predicted ethanol production to increase from 2.6 to 5.7 billion gallons
between 1985 and 2000. Ethanol production became less optimistic in
1981, however, due to a decrease in petroleum consumption, the prospects
of converting cellulose to ethanol, and the increasing use of methane
from coal gasification (Abel, 1982).

The NALS Final Report (1981) predicted 12 million new households to
be added to nonmetropolitan areas between 1977 and 1995, with a 62 to 75
percent increase in the suburban population by 2000. Brown and Beale

(1981) predicted a sharp decline in the rate of household growth after

25
1985, with 2.5 million fewer households in 1995 than if the 1960 and
1970 rates persisted. Heady (1982) stated that if past trends in
nonfarm uses of land continue to 2000, urban areas will occupy about 2
percent of the total U.S. land area - compared with 1.5 percent in
1969.

Many people assumed that the 1967-75 annual conversion rate of 2.9
million acres would continue to 2005 and take 80 million acres out of
agriculture (Crosson, 1982a). But household formation is expected to
decrease after 1985 (compared to the 1967 to 1975 rate) because adults
from the post World War II baby boom will be getting out of household
formation age (Brown and Beale, 1981; Fischel, 1982; Brewer and Boxley,
1982; Dovring et al., 1982; Heady, 1982). Also, the construction of
roads, highways, and airports has slowed and Brown and Beale (1981) did
not anticipate any significant expansions through 2000. Dovring et al.
(1982) noted that low interest rates in the 1970's encouraged
construction of homes, and in Illinois most of the urbanization was due
to rising affluence rather than population growth. The high interest
rates in the 1980's however, dampened construction of new homes. Heady
(1982) believed that the increasing use of multi-unit buildings, slower
population growth, and higher energy prices will dampen urban,

transportation, and recreational demands for land.

CONCLUSION
Urbanization has affected farmland for decades. However, the loss
of farmland did not become an issue until the USDA showed an increase in

the number of acres harvested and, at the same time, the Soil

26

Conservation Service reported an increase in the conversion rates of
agricultural land (Healy, 1982). This may suggest that the agricultural
land retention issue depends heavily on the data base used those who
collect and interpret the data may determine whether an issue develops.
Some people obstensibly concerned about decreased food production,
however, may primarily be concerned with the loss of amenity values -
such as open spaces, and the preservation of rural life styles. Brewer
(Easterbrook, 1986) noted that "when the false arguments about food
production and urbanization are peeled away, what people are really
saying is that they like to live in a rural setting and they don't want
other people coming in." Concerns over the ability to feed a hungry
world were voiced because they would draw political responses (Crosson,
1982a; Healy, 1982).

Land use data have primarily been examined from an agricultural
point of view, with other land uses considered only as they infringe
upon agriculture. There are different perceptions of land use between
city planners and agricultural experts, especially at the urban - rural)
fringe. An agriculturist would probably call this area "urban".
Conversely, a city planner divides urban land uses into many categories-
with all residual areas being "rural" (Clawson, 1982).

The significance of farmland conversion depends on many complex)
factors, such as the quality and quantity of land in reserve, the cost
of converting this land into production, possible new yield-increasing
technologies, future demands for land and food, and the actual I
conversion rates (Sampson, 1981). Crosson (1982a) contended that the /

ability to produce food and fiber was not constrained by physical limits

27
of the land because production can be increased at some cost.
Therefore, the focus should be on products, not just acres (Healy,
1982).

The conversion of agricultural land as a resource problem is not a
present day "crisis" (NALS, 1981), but it does pose some serious long -
term risks to the resource base and food production. The NALS (1981)
report stated that a resource problem was developing, but immediate
incentives for conserving the resource were weak. Fischel (1982)
believed that the NALS painted a pessimistic picture - - one that should
be qualified; the problem is in the future (allowing ample time to
correct it), data showing agricultural land conversion were not
reliable, and factors that caused population decentralization were
weakening.

Proponents of land preservation frequently cited the 3 million
acres per year figure reported by the SCS "because it was impressive,
ominous, and nicely rounded" (Easterbrook, 1986). The NALS assumed that
this value was correct before work began and it was designed to support
the premise that the United States was running out of land (Easterbrook,
1986). Economists Julian Simon, Michael Brewer, Robert Boxley, and
William Fischel disputed the statistical claims of the NALS. They
want more and better data before enumerating policies (Wheeler and
Harper, 1983). Furthermore, many farmers suffering from low commodity
prices conclude there is too much, not too little, farmland (Wheeler and
Harper 1983). Even Associate Secretary of Agriculture John Crowell was
doubtful (Little, 1984). Crowell saw no need to protect farmland

because of present crop surpluses and the promise of new technologies to

28
improve yields. Wheeler and Harper (1983) believed there was too much
blind faith in technology. They conluded that the issue of agricultural
land conversion was plagued with uncertainty. The only way to overcome

this uncertainty is with good data on the conversion of agricultural

lands.

STUDY AREA AND METHODS

The study area included the 40 southern-most counties in Michigan
(Figure 1). This area contained 86.4 percent of the state's cropland
acreage, 90.3 percent of the prime farmland acreage in cropland and
pasture (U.S.D.A., 1984), and 91 percent of the state's population (U.S.
Dept. of Commerce, 1982).

The study area was divided into six regions (Figure 1) based on
population distribution and soil parent materials. Population
distribution was the most important consideration because of its
influence on the sampling rate (Figure 2). Population characteristics
were obtained from the 1980 Census of the Population (U.S. Dept. of
Commerce, 1982). Region 1 consisted of small towns with less than
5,000 people on lake plain soils with some moraines and till plains.
The soils were primarily loamy. Region 2 contained the state's second
most populated counties, with cities ranging from 2,000 to, 80,000
people. The soils were loams and loams over sand and gravel. The
landforms consisted of outwash plains, till plains, moraines, and lake
plains. Region 3 contained the state's most populated counties and
largest metropolitan areas. The soils were mainly loams or loams over
sand and gravel. The landforms were lake plains, moraines, outwash
plains, and till plains. Region 4 consisted of cities of less than

5,000 people, with each county containing a city of 6,000 to 10,000

29

 

Figure 1. Study Area

30

       

 

 

 

I
I
‘ O
‘ a
\ \ u. :3
\
“xx‘ \ :3
I o \ \
I D \\ x 3
I \\ w a
I \ V "(sour ls”
'I Q C".“' 075‘60 MO~7H°.. AL" NA
I ANFRIH
o 0
I
\
‘ ”(UM JMH‘SK- Cfian'r‘o oscou “cam:
”'0 row.

stun:
sun's. aria/000 MIJJAUa: NOJCUM. oat-Mar rosco

 

 

 

 

 

 

 

LAN! OSCIEOLA CLAMP! GIADN. AR‘NAC

   
  

 

Nl narcoiutcos IA asap“ LA MIDAAN

 

 

 

 

 

 

 

 

 

 

 

 

x uunou
. . 3 "Atom .smvuAC
MONICALM Gn’naor ‘ .
—
«(NT . ‘
Iv :.
. IOIVIA CLIN70N NIAOVA. Q S T. CLAIR
h-
. . . t o
1
OAIHAND
‘lVl
o

 

”0157‘” W

. §'
Detroit
bkANCn n’iusoa. LEIquvl't Dumbo: ‘ Metrogolitan
o O 0 area

 

    

 

 

 

O

 

 

 

Figure 2. Study Area With Cities

31

32

people. The soils were loams and loams underlain by sand and gravel.

The landforms were till plains, outwash plains, and moraines. Region 5
had the same population characteristics as Region 4. The soils in the
northern and western portions of the region were sandy, and the soils in
the southern and eastern portions were loamy. The landforms were
primarily outwash plains, till plains, lake plains, and moraines.
Region 6 had pOpulation characteristics similar to Regions 4 and 5, but
each county contained a metropolitan area with 50,000 to 180,000 people.
The soils were much sandier than those in Region 2, with some loams
underlain by sand and gravel. The landforms consisted of moraines,

outwash plains, and sand dunes.

Sampling
A stratified random sampling technique (Cochran, 1963; Kish, 1965)
was employed using a section of land (640 acres) as the sampling unit.
Stratification helped reduce the sample size (for a given level of
error) by reducing the variability between samples. Each county was
subdivided into strata using 1980 Census of the Population data (U.S.
Dept. of Commerce. 1982). The strata consisted of:
Stratum 1: Cities < 5,000 inhabitants
Stratum 2: Cities Z 5,000 inhabitants
Stratum 3: Townships < 10,000 inhabitants
Stratum 4: Townships 2,10,000 inhabitants
with cities always being stratified first.
Cities and townships identified in the 1980 census were delineated on
soil survey "index to map sheets" for each county (this made it easier

to stratify and select samples). Stratum 1 consisted of sections

containing town boundaries. Stratum 2 consisted of all sections within

33
and adjacent to the city boundaries. Stratum 3 consisted of townships
with less than 10,000 people. Stratum 4 consisted of 1) highly
populated townships, 2) sections outside of large cities that were
influenced by the city but not enough to be included in that stratum,
and 3) large developments (g_30 acres) such as highways, gravel pits,
and land fills in rural areas.

Samples were selected in two steps. The first step was to
consecutively number all strata in a region and randomly select a
representative number of samples. The second step was to randomly
select a section from each sample chosen previously. If a city or
township was selected more than once during the first step, then the
appropriate number of sections were selected in step two. Large lakes
of Z 160 acres, and inner city areas were excluded from the sampling
frame (areas that did not contain any agricultural uses in 1962).

Developed areas in 19621 were identified using Agricultural
Stabilization and Conservation Service (ASCS) 1:20,000 or 1:15,840 scale
black and white photographs. These areas were delineated on a soil
survey map sheet. Michigan Department of Natural Resources (MDNR) 1978
color infrared (CIR) photographs were used to identify developments
since 1962 . These areas were also delineated on the soil survey map
sheet. The CIR photographs made it easier to detect land uses and thus
1 served as a check for 19621 interpretations. An attempt was made to
use 1962 imagery for each county in a region, and every region, (except
Region 5 and Wayne County in Region 3) had imagery dates within + or
- 1 year (see Appendix A).

1
Dates varied between 1958 and 1964 (see Appendix A)

34

Developed uses consisted of buildings (houses,
businesses, factories, etc.), highways, mines, gravel pits, borrow pits,
water bodies produced by digging (not flooding), parks, oil refineries,
cut and filled areas, airport runways and buildings, trailor parks, golf
course buildings, and other types of development that either paved over
and compacted the soil, or left materials in the ground, thus making it
too costly to convert to cropland. Water bodies produced by flooding
were not included as developed areas because they could be drained.
Parks were included as developed areas because they usually contained
water lines, sewer lines, and electrical hook-ups.

Developed areas were delineated using square or rectangular

boundaries unless a distinct curvilinear boundary was present. Yards
were included with homes; highway right-of-ways were included with
highways; and disturbed areas adjacent to new structures were included
with the structures. Some assumptions were made in identifying
developed uses, they included:
1. When sampling around cities, open spaces were considered to be
developed uses unless agriculture was clearly the predominant use. To be.
concidered an agricultural use, open spaces could not be bordered on all
sides by developed uses, or they had to show evidence of an agricultural
use, and adjacent areas had to be in an agricultural use.

2. When sampling rural areas, all open spaces were given agricultural

priority.
3. Only current land uses were identified, not intended uses.
4. Land uses were identified as "developed" or "open." Land use

classifications were not made.

35
5. Areas classified in the soil survey as "made land," or "fill"
were classified as developed.
6. The study considered prime soils rather than prime farmland. By

definition, prime farmland must be in an agricultural or forestry use.

After delineating the developed areas, they were measured using
a Lasico Polar Compensating Planimeter. Developed areas in 19621 and
1978 were recorded along with development on prime soils, good soils,
and "other" soils. Any farmsteads destroyed since 19621that had yards
annexed to a field were subtracted from development totals. Prime and
good soils were identified from lists provided by the Soil Conservation
Service (Hilner, 1983; Hilner, 1985), and drainage was assumed for prime
soils with a drainage requirement.

Total development since 19621 along with total development on
prime, good, and other soils was determined using stratified random
sampling statistical techniques (Cochran, 1963; Kish, 1965) and the
computer program STRTRAN (Barrett and Nutt, 1975). The sample size was
determined using optimum allocation (Kish, 1965; Barrett and Nutt, 1975)
and the computer program STRTSIZE (Barrett and Nutt, 1975).

Determination of the sampling error was based on the amount of
development between 19621 and 1978. An acceptable sampling error was 10-
15 percent with 95 percent confidence. The sampling error was
calculated by

2[S(YW)]/YW * 100

where
s(yw) is the total sample standard error of the mean

yw is the total sample mean

36
Three counties were selected to determine the rates of development
between 1950 and 1978. Bay, Berrien, and Huron Counties were
selected because imagery was available and they were considered to
represent all regions. Huron County represented region 1; Berrien
County represented Regions 4, 5, and 6; and Bay County represented

region 2, and region 3 (but on a smaller scale).

1
Dates varied between 1958 and 1960 (see Appendix A)

RESULTS AND DISCUSSION

Regional totals of agricultural land conversion within the study
area varied from 10,614 acres in region 1 to 187,282 acres in region 3
(Table 4). These values were determined at 95 percent confidence with a
sampling error of 10-15 percent, and show a direct relationship between

the amount of development and the population of the regions.

1
Table 4 Total Conversion Of Agricultural Land Between 1962 - 1978

------------------- Acres-----------------
Region *Total Area Development Confidence Limits Population T\§
*'\

ix

1 2,463,144 10,614 8,591-12,637 343,049 ”\
2 2,772,205 95,890 83,003-108,777 1,274,651
3 1,388,726 187,282 158,530-216,053 4,494,733
4 2,789,664 30,074 25,500-34,648 426,158
5 4,531,223 64,352 54,763-73,942 622,537
6 1,912,744 98,770 85,006-112,533 1,142,923

In predominately rural areas (regions 1, 4, and 5) the largest
source of farmland conversion was from cities of 2_5,000 people. These
areas accounted for 41 percent of the conversion in region 1, 37 percent
in region 4, and 44 percent in region 5. The second largest source of

farmland conversion in these regions was due to low density housing on

1
Dates vary between 1958-1964 (see Appendix A).

* Total area only included land areas mapped by SCS; therefore, lakes,
urban areas, and pits are not included.

37

38
unsewered lots in areas away from towns (townships with less than 10,000
people). This type of development accounted for 37 percent of the
conversion in region 1, 35 percent in region 4, and 41 percent in region

5. These results are consistent with those reported by Rossol (1982).

Table 5 Farmland Conversion By Population Distribution (Acres)
------------------ Population------------------
Cities Cities TownshipsTownships
Region < 5,000 3 5,000 < 10,000 2_10,000
1 1,435 4,275 3,970 934
2 5,731 44,123 16,417 29,619
3 5,450 136,535 8,033 37,274
4 5,753 10,998 10,486 2,837
5 5,602 28,566 26,342 3,842
6 8,507 59,520 14,895 15,848

Conversion of prime farmland varied from 4,572 acres in region 1
to 81,847 acres in region 3 (Table 6). The conversion of good and other
farmland varied from 3,909 and 2,133 (respectively) in region 1 to
63,091 to 42,254 acres (respectively) in region 3. These values
indicate a direct relationship between the amount of development on
prime, good, and other soils and the population of the regions. Between
19621 and 1978, approximately 231,500 acres of prime soils,* 133,700
acres of good soils, and 121,900 acres of other soils were converted to
non-agricultural uses in the study area. The sampling error for
development on prime, good, and other soils varied more than for total
development (as indicated by the confidence limits in Table 6) because

1

the sampling rate was based on total development since 1962 . The 1982

1
Dates varied between 1964 and 1958 (see Appendix A)

39
NRI (Misseldine and Riggle, 1985) reported a loss of 600,000 acres of
prime farmland between 1977 and 1982.

1
231,500 acres converted between 1962 and 1978 in the study area.

This value is much higher than
The
NRI estimate was for the whole state of Michigan, but the sample area in
this study contained 91 percent of the state's population (U.S. Dept. of
1982) and 90.3 percent of the state's prime farmland acreage

Commerce,

used for cropland and pasture (U.S.D.A., 1984).

Table 6 TOTAL LAND CONVERSION BY SOILS BETWEEN 1962 - 1978
--------------------------- ACRES-----------------------
Region Prime Good Other
1 4,572 3,909 2,133
2 54,025 25,217 16,648
3 81,947 63,091 42,254
4 16,786 4,520 8,768
5 37,840 13,345 13,307
6 36,318 23,621 38,803
Confidence Limits
Region Prime Good Other
1 2,929 - 6,222 2,639 - 5,180 1,178 - 3,095
2 42,834 - 65,216 14,948 - 36,199 8,738 - 25,272
3 53,450 109,687 39,615 - 85,809 17,232 - 59,276
4 13,477 20,096 2,306 - 6,377 5,425 - 12,111
5 30,062 45,619 8,136 - 18,504 6,826 - 19,697
6 25,911 46,725 14,309 - 32,948 25,510 - 52,108
Prime soils were developed more often than good and other soils

(Tables 6 and 7). This was partly due to a greater proportion of prime

soils to good and other soils (Table 8). In regions 2, 4, and 6

conversion was proportionate to the extent of the three groups of soils

(Tables 7 and 8).

1
Dates varied between 1958-64 (see Appendix A)

40

Table 7 Percentage Of Development By Soils

Region 2 Prime 2 Good 2 Other
1 43 37 20
2 56 26 17
3 44 34 23
4 56 15 29
5 59 21 21
6 37 24 39

Table 8 Distribution Of Prime, Good, and Other Soils In Each Region

Region 2 Prime 2 Good 2 Other
1 62 29 9
2 57 31 12
3 58 35 7
4 60 22 18
5 41 25 34
6 39 27 34

Less than 3.5 percent of the prime soils in regions 1, 2, 4, and S
were developed during the twenty six - year period (Table 9).
Approximately 5 percent of the prime soils in region 6 and 10 percent of
the prime soils in region 3 were developed during this period. The
percentage for region 3 was lower than expected considering the region
contained 54 percent of the population in the study area and the fifth
largest metropolitan area in the nation, Detroit (4,361,000 people)
(Morss, 1986). The conversion of good soils to developed uses was the
highest in region 3 (13 percent) and the lowest in region 1 (0.6
percent). Good soils in region 6 experienced approximately the same
amount of development as prime soils in the region (Table 9). "Other"
soils in regions 2 and 6 experienced more development than prime and
good soils in the respective regions. In regions 1, 3, 4, and 5

development on other soils was less than 2 percent (Table 9).

41

Table 9 Percentage Of Prime, Good, and Other Soils Developed

Region 2 Prime 1 Good Z Other
1 0.3 0.6 0.9
2 3.4 3.0 5.1
3 10.3 13.0 0.4
4 1.0 0.8 1.7
5 2.0 1.1 0.9
6 4.8 4.6 6.0

Rates of farmland conversion in the study area were measured over
different time intervals in three representative counties (Huron,
Berrien, and Bay) (Figures 3, 4, 5 and 6). All three counties had a
dramatic increase in development during the 1960's, and two of the
'counties (Huron and Berrien) showed a decrease in the annual rate of
development in the 1970‘s (Figure 6). This trend is similar to that
reported by the Census of Agriculture, ESS, NALS, and SCS (see Table 2).
Imagery was not available for Bay County in the early 1970's;
therefore, any decrease in the early 1970's has been averaged in with
the increase in the 1960's. Conversion rates in region 1 are expected to
be similar to those in Huron County. Conversion rates in regions 4, 5,
and 6 are expected to be similar to those in Berrien County.
Conversion rates in regions 2 and 3 are expected to be similar to those
in Bay County, but conversion in region 3 would be of greater magnitude.
Much of the increase in conversion during the 1960's was due to
construction of the interstate highway system. By the fall of 1986,
however, the interstate highway system will be completed in Michigan
(Adams and Bailey, 1986).

There was a direct relationship between the amount of

development and the population of the regions. In predominately rural

mv

JLCDDU CDLJI r; C0_mtm>CDU UCEELmu .m mtjmi

mmamn .
mhINNE Nulwmg vmnnmm. hmlomm_ amt—ww—

[eaJov 0m) BBJV pedolaaag

 

 

1 987-78

|IIIIIIIIIIIIIIIIIIIIIIIIII

_—_———.—_——_—————.—o———_.

     
 

1* .__.» _. u.- - 2‘ _ _— _.- _— _. _+ _ ~— .— ———- -__

._—_—————‘———_——-—-.-_fi

I"!!!"'! I

 
  

1 380-87

 

l 955-80

 

1 950-55

 

l I I l l l
(D i\ (D ID V’ (D N -- D

 

I I

[BSJDV 00m) BBJV padolaAaQ

Dates

Figure 4. Farmland Conversion in Berrien Countg

43

             

‘eeeee
/ o

9’ \ / . OK. ‘9‘ 97 73‘ 9, OT '9‘ 9
”\o/ . \./ O / / 4 /. \4

. -0 /9x
./ \ \ ./ \/ \

06 oe/\/o\%’/\’/%®/%‘®’”~/
>®/\/\/.\/\/\/.\/.\/.\ .\/\/\/.\/\/\./.\.1
/\/\/\/\/\/\ \/\/ ./\ \/ / ‘ /’“
to e o 0 ©®€®©©Q

/A\ /‘\ /A‘~ /A\ /A‘. /A‘ /A. /‘.x /A /A /A‘a /A\ /‘ /A\ /‘ /‘ /.‘

   
   

 

           

    

 

       

 

[BBJOV 00m] BBJV petiole/mu

 

1 983-78

1 950-83

Dates

Figure 5. Farmland Conversion in Bag Countg

44

me

mm$CDDU 3mm Dcm .Cmttmm 60.51
E mmwmm C0_mtm>CDU UtmEtmn. _m_.._r_r_< .m 0.59“—

unann—

Dm on am am on

 

 

_ e _ _ _ _ . U
cotnzi _ m
m.

I N m

l m o.

.. . w

mam 1 m m.

l m U

cmmttmm l n m
I m W

J

l m m

 

46

areas (regions 1, 4, and 5) however, 35 to 41 percent of the farmland
conversion came from low density housing on unsewered lots away from
cities. Ten percent of the prime soils and 13 percent of the good soils
in Region 3 (containing Flint and metropolitan Detroit) were developed
since 19621, the highest in the study area. With low amounts of prime

and good soil development, the greatest amount of prime soil
development adjacent to cities, a downward trend in the conversion rates
of agricultural land, and 12,136,452 acres of prime and good soils
undeveloped in the study area, it would be impractical to pass
legislation protecting all prime soils, especially with 91 percent of
the state's population (U.S. Dept. of Commerce, 1982) interspersed with
90.3 percent (U.S.D.A., 1982) of the state's prime soils used for
cropland and pasture. If prime soils are to be protected, the focus
should be upon low density housing on unsewered lots remote from towns.

It should be noted, however, that even though the emphasis is on prime

and good soils, other soils are used for crop production.

1
Dates varied between 1964 and 1958 (see Appendix A)

CONCLUSION

It is possible to objectively and inexpensively collect and update
data on farmland conversion using aerial photograph interpretation.
This technique can be used at the local, regional, or state level.
About 487,000 acres of farmland were converted to nonagricultural uses
between 1962 and 1978, significantly less than earlier estimates. This
,is only 3 percent of the prime, and 3 percent of the good soils in the
study area. Prime soils experienced more development than good and
other soils. This was partly due to a predominance of prime soils. With
large acreages of prime and good soils in the study area, coupled with a
large population, slowing conversion rates, and small percentages of
prime and good soil development, it would not be advisable to adopt
legislation preserving farmland. Rural regions within the study area
experienced large conversion amounts from low density housing on
unsewered lots, remote from any towns. If farmlands are to be

protected, however, this is where policy decisions should be targeted.

47

FUTURE RESEARCH RECOMMENDATIONS

There are three areas where additions could be made to the
research. First, obtain the most recent imagery available to detect any
changes in the amount and rate of agricultural land conversion. The
sample size could be reduced by treating the study area as one
population rather than six. To do this, imagery within a four year
period would be needed for each county in the area.

Second, a stratification technique that takes into account 'soil
properties‘ could be developed to reduce the variability on the total
conversion for each soil. If this is not possible, an area that
contains a combination of the three soils should be sampled more heavily
than an area dominated by one soil.

Third, analyze the data from both an economic and physical
perspective. The economic analysis should focus upon both supply and

demand for agricultural land.

48

LIST OF REFERENCES

Abel,

Adams,

Barrett,

Boxley,

Brewer,

Brewer,

Brown,

Caughlin,

Clark,

Clawson,

Cochran,

LIST OF REFERENCES

M.E. 1982. Growth in Demand for U.S. Food and Animal
Production by 2000. The Cropland Crisis: Myth or
Reality, edited by P.R. Crosson. John Hopkins University

Press, Baltimore. p. 63-88.

G.R. and E.W. Bailey. 1986. Michigan Dept. of Transportation.
Lecture. Lansing, MI.

J.P. and M. Nutt. 1975.
Sciences: A Computer Approach.
Hanover, N.H.

Survey Sampling in the environmental
Compute, Dartmouth College.

R. 1981. Competition for Agricultural Land to the Year 2000.
Agricultural Land Availability, edited by M.F. Brewer.
U.S. Government Printing Office, Washington, D.C. p. 103-186.
M.F. 1981. Agricultural Land Availablity. U.S. Government
Printing Office, Washington, D.C. p. 3-10.

M. F. and R. Boxley. 1982. The Potential Supply of Cropland.
The Cropland Crisis: Myth or Reality, edited by P.R.
Crosson. John Hopkins University Press, Baltimore. p. 93-109.

D.L. and C.L. Beale. 1981. Sociodemographic Influences on
Land Use in Non-MetrOpolitan America. Agricultural Land
Availability, edited by M.F Brewer. U.S. Government Printing
Office, Washington, D.C. p. 77-102.

M. 1981. Adequacy of Land Use Information.
Land Availability, edited by M.F. Brewer.
Printing Office, Washington, D.C. p. 269-282.

Agricultural
U.S. Government

J. 1979. Conserving the Nation's Farmland: A Background Paper.
Northeast-Midwest Institute, Washington, D.C. p. 1-7.

M. 1982. Discussion. The Cropland Crisis; Myth of Reality,
edited by P.R. Crosson. John Hopkins University Press,
Baltimore. p. 110-112.

W.G. 1963. Sampling Techniques. 2nd Ed. Wiley and Sons, Inc.
New York.

49

Crosson,

Crosson,

Dovring,

50

P.R. 1982a. The Long-Term Adequacy of Agricultural Land in
the United States. The Cropland Crisis: Myth or Reality,
edited by P.R. Crosson. John Hopkins University Press,
Baltimore p. 1-22.

P.R. 1982b. Future Economic and Environmental Costs of
Agricultural Land. The Cropland Crisis: Myth or Reality,
edited by P.R. Crosson. John Hopkins University Press,
Baltimore. p. 165-191.

F.D., D.L. Chicoine, and J.B. Braden. 1982.
Evaluating Agricultural Land Use Changes In Illinois. J. Soil
Water Conserv. 37:359-361.

Easterbrook, G. 1986. Vanishing Land Reappears. The Atlantic. July,

Fischel,

Heady,

Healy,

Hilner,

Hilner,

Huffman,

Jeffords,

Kish,

Little,

McGill,

1986. pp.17-20.

W.A. 1982. The Urbanization of Agricultural Land: A Review of
the National Agricultural Lands Study. Land Economics 58:236-
259.

E.O. 1982. The Adequacy of Agricultural Land: A Demand-Supply
Perspective. The Croland Crisis" Myth or Reality, edited by
P.R.Crosson.John Hopkins University Press, Baltimore. p.23-56.

R.G. 1982. Discussion. The Cropland Crisis: Myth or Reality,
edited by P.R. Crosson. John Hopkins University Press,
Baltimore. p. 113-116.

H.R. 1983. Soils To Be Considered in the Selection of
Farmlands of Local Importance. U.S.D.A. S.C.S. East Lansing,
Michigan.

H.R. 1985. Prime Farmland map Units in Michigan. Tech. Guide
49. U.S.D.A.S.C.S. East Lansing, MI.

B. 1981. Agricultural Land in National and Regional Economies.
Agricultural Land Availability, edited by M.F. Brewer. U.S.
Government Printing Office, Washington, D.C. p. 11-76.

J.M. 1979. Protecting Farmland: Minimizing the Federal Role.
J. Soil Water Cons. 34:158.

L. 1965. Survey Sampling. Wiley and Sons, Inc. New York.

C.E. 1984. Farmland Preservation: Playing Political Hardball.
J. Soil Water Cons. 39:248-250.

E. A. Hidlebaugh, and J. Yovino. 1981. Federal Data on
Agricultural Land Use. Agricultural Land Availability,
edited by M.F. Brewer. U.S. Government Printing Office,
Washington, D.C. p. 239-268.

Michigan

M.D.A.

M.D.A.

Misseldine,

Morss,

National

Raup,

Reganold,

Rossol,

Sampson,

Supalla,

U.S.D.A.

U.S.D.A.

U.S.D.A.

U.S.

Wheeler,

51

Department of Agriculture (M.D.A).1959. Michigan Agricultural
Statistics. Lansing, MI. pp. 14.

1976. Michigan Agricultural Statistics. Lansing, MI.
1978. Michigan Agricultural Statistics. Lansing, MI.

Michigan's Land and Soil
East

C. and J.D. Riggle. 1985.
Resources: A Status Report. American Farmland Trust.
Lansing, MI.

D. 1986. Michigan Travel Bureau.
the 1986 Survey of Buying Power.
Management Magazine. Lansing, MI. -

Providing Information from
Sales and Marketing

Agricultural Lands Study NALS. 1981. Final Report. U.S.
Government Printing Office, Washington, D.C. p. 8-60.

P.M. 1982. An Agricultural Critique of the National
Agricultural Lands Study. Land Economics 58:260-274.

J.P. 1986. Prime Agricultural Land Protection: Washington
State's Experience. J. Soil Water Con. 41:89-92.

M. 1982. Urbanization and its Relationship to Prime Farmland
in Two Michigan Counties. M.S. Thesis. Michigan State Univer-
sity. East Lansing, MI.

Rodale

R.N. 1981. Farmland or Wasteland: A Time To Choose.

Press, Pennsylvania.

R.J., R.R. Lansford, and N.R. Gollehon. 1982. Is the Ogallala
Going Dry? J. Soil Water Cons. 37:310-314.

Soil Conservation Service. 1968. Michigan Conservation Needs
Inventory. East Lansing, MI.

Soil Conservation Service. 1982. Basic Statistics: 1977
National Resources Inventory. Statistical Bull. 686. U.S.
Government Printing Office. Washington, D.C.

Soil Conservation Service. 1984. Michigan Data: 1982 National
Resources Inventory. East Lansing, MI.

Department of Commerce Bureau of the Census. 1982. Number of
Inhabitants: Michigan. U.S. Government Printing Office.
Washington, D.C.

D.P. and S.F. Harper. 1983. In Defense of Farmland. J. Soil
Water Cons. 38:4.

52

Whiteside, E.P. and D. Schaner. 1972. Michigan Agriculture: Agricultural

Wright,

Wright,

Wright,

Zeimetz,

Land Use Trends and Future Needs for Agricultural Land.
Michigan State University, Michigan. 21 pp.

K.T. and D.A. Caul. 1967. Michigan's Agriculture: Its Income,
Major Products, Locations and Changes. Mich. State Un. Ext.
Bull. 582. East Lansing, MI.

K.T. 1974. Michigan's Agriculture: Location and Changes in
Major Products, Number of Farms and Income. Mich. State
Un. Ext. Bull. 785. East Lansing, MI.

K.T. 1984. Michigan Agriculture:1982 County Data. Mich. State
Un. Ext. Bull. East Lansing, MI.

K.A., E. Dillon, E.E. Hardy, and R.G. Otte. 1976. Dynamics of
Land Use in Fast Growth Areas. Agr. Econ. Rep. No. 325. USDA
Econ. Res. Serv., Washington, D.C.

APPENDICES

APPENDIX A

Table 10 County Data Sources

County

Huron
Lapeer
Sanilac
St. Clair
Tuscola
Bay
Ingham
Jackson
Livingston
Monroe
Saginaw
Washtenaw
Genesee
Macomb
Oakland
Wayne
Allegan
Branch
Cass
Hillsdale
Lenawee
St. Joseph
Van Buren
Barry
Clinton
Eaton
Gratiot
Ionia
Isabella
Mecosta
Midland
Montcalm
Newaygo
Oceana
Shiawassee
Berrien

Region

O‘U‘UIU'IU'IU‘MUIMUIU‘UIUIbkbkbbbwwwwNNNNNNNi-‘HHHH

Date

1964
1964
1963
1964
1963
1963
1963
1964
1963
1964
1963
1963
1964
1964
1964
1957
1960
1961
1960
1963
1963
1960
1960
1962
1963
1960
1963
1960
1958
1958
1958
1961
1958
1958
1963
1960

Source

P

"U’U'U'U'U*d’d'd'd'd'd'd’d’d’dz'd'd'fl'fi'd"U’UUJZZ'UZ'US'U"U'UZUI

53

Imagery Imagery Number of

Samples

25
11
12
18
11
15
10
13

5

6

9

7
17
16

moobmubwosmoooua

Hp

1980

Popoulation

36,459
70,038
40,789
138,802
56,961
119,881
275,520
151,495
100,289
134,659
228,059
264,748
450,449
694,600
1,011,793
2,337,891
81,555
40,188
49,499
42,071
89,948
56,083
66,814
45,781
55,893
88,337
40,448
51,815
54,110
36,961
73,578
47,555
34,917
22,002
71,140
171,276

Soils
Information

Survey
Survey
Survey
Survey
Interim
Survey
Survey
Survey
Survey
Survey
Update
Survey
Survey
Survey
Survey
Survey
Interim
Interim
*Interim
Update
Survey
Survey
Interim
*Interim
Survey
Survey
Survey
Survey
Survey
Survey
Survey
Survey
Update
Update
Survey
Survey

Appendix A (cont'd)

Kalamazoo 6 1960
Kent 6 1961
Muskegon 6 1962
Ottawa 6 1962

P Photograph

S Soil Survey

M Photograph Mosaic (Index)
*

"U’d’v'd

54

14
14

10

These counties were partially mapped

212,378
444,506
157,589
157,174

Survey
Interim
Survey
Survey

Interim Reports refer to advanced soil map sheets for counties being

mapped

Table 11

APPENDIX B

Soil Acreages by Region and County

County

Huron
Lapeer
Sanilac
St. Clair
Tuscola

County

Bay
Ingham
Jackson
Livingston
Monroe
Saginaw
Washtenaw

County

Genesee
Macomb
Oakland
Wayne

REGION
Prime

348,665
227,344
429,742
270,143
251,064

REGION
Prime

194,845
251,625
190,870
135,445
260,009
338,752
225,350

REGION
Prime
286,936
216,401

199,830
95,740

55

1
Good

84,795
154,789
160,870
179,279
128,249

Good

49,875
94,825
222,085
173,409
55,545
71,168
183,490

3

Good

Other

94,140
36,347
24,061
22,099
51,557

Other

39,810
8,560
38,030
55,804
38,260
107,328
37,120

Other

82,583
51,329
245,280
106,125

22,453
25,904
43,505
12,640

"IIIIIIILIIIIIIIIIIIIIIIIII