Author: Humberto G. dos SantosPublisher:
ISBN:
Category : Soil surveys
Languages : en
Pages : 302
Book Description
Soil maps of nine areas representing distinct physiographic patterns and land use were evaluated by three different procedures to study the accuracy in predicting mapping unit composition, and the relevance of procedures for different kinds of areas in relation to effort required. The physiographic and soil patterns studied are believed to be representative of range, forest, and mixed range and forest in mountainous terrain, nearly level, rainfed and irrigated cropland, general farming in rolling and hilly terrain, coastal areas and desert areas. Delineations of selected soil association on generalized soil maps at scales ranging from 1:190,080 to 1:380,160 were tested by point transects, line transects and pilot areas using soil surveys at scales of 1:20,000 or larger as the ground truth. Topographic maps and aerial photographs were used to identify physiographic units, land use patterns and topographic characteristics. Random transects that crossed the grain of the landscape were selected and drawn on detailed soil maps of each of the study areas. With the point transect method, soil units were noted on pre-fixed point at constant intervals of 0.05 kilometer up to 4.0 kilometers along the paths of observation. Soil units were noted along the same transect using different intervals, that is, observations at each 0.05 kilometer along its entire length, then at each 0.1 kilometer and so on up to 4.0-kilometer intervals. Line transects were located similarly, and all soil changes along the of observation noted. Pilot areas were selected to depict small, representative areas according to physiographic features that could be extrapolated to similar units of landscapes. Point transects provided the number of point observations of each soil unit; line transects provided length, in kilometers, of soil units between soil changes; and pilot areas provided the extension, in hectares, of each soil unit observed. Because transects and pilot areas were of different sizes, data obtained by these procedures were statistically analyzed by a cluster type of analysis to estimate proportions of kinds of soils in each association. In additon to proportions of soil units, confidence intervals, accuracy of predicting mapping unit composition, standard deviation, standard error of the mean, coefficient of variation and number of elements needed (transects or pilot areas) were calculated. The numbers of transects and pilot areas were calculated to meet certain probability levels. For example, 80 percent probability with 10 percent allowable error around the mean was used for mountainous areas, and 95 percent probability with 10 percent allowable error for all other areas. Point transects with intervals between observations ranging from 0.05 Kilometer to 0.2 kilometer required more effort per square kilometer than line transects or point transects with larger intervals between observations. Effort per square kilometer decreased with increasing spacing between observations along point transects, but the number of transects needed to meet predetermined levels of accuracy tended to increase when intervals larger than 0.5 kilometer were used. Increasing the number of transects or pilot areas to test soil association provided more accurate estimates of soil proportions and smaller erros of estimations, however, with increasing effort per square kilometer. Based on number of transects and pilot areas needed to attain the accuracy level fixed for the study areas, it was found that the effort per square kilometer varies with profound that the effort per square kilometer varies with procedure and type of landscape. In general, the point and line transects required less effort per square kilometer in all areas tested. Point transects with intervals between observations ranging from 0.2 kilometer to 2.0 kilometer generally required less effort in most of the test areas. For some areas the line transect method is equivalent or slightly better than point transects in terms of effort in man-hours/km2.
