- A study was undertaken in the Willamette Valley, Oregon, to document
1971 land use patterns as well as land use change in a smaller study
area in the 16 years prior, and to relate these to soil characteristics.
Quantification of these inter-relationships was aided by use of computer
tabulation and graphic methods. Use of the data base generated and
methods developed was expanded somewhat and applied to calculation
and mapping of composite indices of suitabilities and limitations of
soils for selected uses.
Mapping of land use for 1971 and 1955 was done on color and
black and white photos respectively, having a scale of about 1:62,500.
To avoid creating artificial, apparent land use changes by differences
in delineation and classification, mapping was done first on the higher
quality 1971 color photos and then by comparison on 1955 black and white
photos. Importance of various photo-interpretive factors to legend
development and consistent delineation and classification became
apparent and was documented. The legend utilized was, with some
modification, the 1972 U.S.G.S. proposed national system for use with
remote sensor data.
To facilitate comparison of land use and land use change with
soil survey map information, photo mapped land use information and
available soil association map information was transferred to transparent
mylar film over planimetric U.S.G.S. 15' topographic quadrangle sheets.
This also allowed use of the square, 1000 meter Universal Transverse
Mercator (UTM) grid as reference for point sampling and geo-referenced
coding of land use and soil information for computer storage.
Tabulation of land use classes for the entire study area for 1971
showed the largest proportion of land to be in agricultural (53.5
percent) and forest land (38 percent) use. A further 6.5 percent
was in urban use. Within the selected townships of the land use
change study area the proportions were different; 65.7 percent,
14.3 percent and 17.2 percent for agricultural, forest, and urban
Analysis of changes in land use between 1955 and 1971 in the
land use change study area was done by means of transition matrices;
diagonal elements represent areas of land not subject to change; off
diagonal elements represent components of change. In total, between
55.9 and 72.4 percent of the land in the area studied experienced
no use change, the range being due to uncertainty of classification
of some land on the 1955 black and white photos. Urbanization was
shown to have occurred largely at the expense of agricultural land and particularly non-irrigated cropland, pasture and orchard land
uses. Some conversion of forest to urban land use (0.9 percent
of the area) was documented, however. The most significant
agricultural use change was a net increase in irrigated land of
between 5 and 8.4 percent. A net decrease in orchard land was also
Land use for 1971, not unexpectedly, showed significant concentration
in certain physiographic areas and in areas with particular soil
characteristics. Thus for example, urban uses were found more
frequently than expected on valley-floor and stream-cut terraces,
and forest land uses more frequently on foothills and mountainous
uplands. Irrigated cropland was somewhat concentrated on soils in
sandy, coarse loamy, coarse silty and fine silty particle size classes.
Urban and cropland agriculture use classes have obviously avoided
steeply sloping areas. Irrigated cropland and orchard land uses have
avoided soil areas with high shrink-swell potential whereas grass
seed producing croplands have not. Numerous other examples could
be given. In summary, Chi-square analysis on the contigency tables
of land use class and physical land characteristic classes gave
highly significant values (eg. for 1971 land use/physiographic
² calc = 9108.25 and x
² tab = 137.68), and thus the null
hypotheses of no association between land use classification and
these physical land characteristics were rejected.
Three-way tabulation of soil characteristics and physiographic
area classification against 1971 and 1955 land use classification
allowed analysis of land use change in relation to these physical
properties. The idea of composite model soils was used to provide
a framework for summary of the kinds of soils subject to smallest
and largest changes in particular kinds of urban, agricultural, and
forest land uses. As examples, it appeared that greatest relative
increases of residential and industrial uses occurred on poorly
drained, fine textured, slowly permeable soils; greatest relative
losses of agricultural land occurred on soils having the highest
capability for agriculture, and; greatest relative increases in
irrigation cropland uses occurred on soils rated good to excellent for
irrigation. Other examples and trends were observed.
A method for computer calculation of composite soil suitability
or limitation indices for various uses was developed. Ranked indices
were compared to interpretive groupings by best subjective judgement.
Class frequency correlations were found to be generally good. Indices
were further mapped using a computer graphics system, GRIDS. The
index calculation method interfaced with a mapping capability was
shown to have potential for rapid, flexible display of various soil,
land use, as well as soil and land use patterns for land use