The Santa Rita Experimental Range (SRER) soils are mostly transported alluvial sediments that occur on the piedmont slope flanking the Santa Rita Mountains in Arizona. The major geomorphic land forms are alluvial fans or fan terraces, but there are also areas of residual soils formed on granite and limestone bedrock, basin floor, stream terraces, and flood plains. The soils range in age from recent depositions to soil material one to two million years of age. We sampled A and B horizons of soil series from different geomorphic surfaces, and measured the dry spectral reflectance (0.4 to 2.5 mum wavelength) on the sieved less than 2-mm-size fraction. Soil color (measured with a Chroma Meter), texture, organic carbon, calcium carbonate content, and effervescence properties were determined and correlated to spectral reflectance in selected wavelengths. The Munsell soil color value component was most positively correlated to reflectance. Soil effervescence and calcium carbonate content, percent sand and clay, and the Munsell soil color hue component and redness rating were also significantly correlated to soil reflectance. Energy reflected from soil surfaces represents the interaction between many soil properties, and soil color is an integrated expression of many soil properties. It is the best soil morphology property to measure to predict the spectral reflectance of soils, particularly in the visible and near infrared parts of the electromagnetic spectrum.

Knowledge of how surface roughness influences the reflectance of sunlight from cultivated soils is useful in various applications, such as estimating albedo values used as inputs to soil temperature models and erosion models. The albedos of two soils were studied for dry and wet surfaces with four different roughness conditions, changed from a reference smooth soil surface. The soils were the Gila fine sandy loam [coarse‐loamy, mixed (calcareous), thermic Typic Torrifluvent] (Ap horizon), 10YR 6.0/3.2 dry and 10YR 4.1/3.3 wet, and the Pima clay loam [fine‐silty, mixed (calcareous), thermic Typic Torrifluvent] (Ap horizon), 10YR 5.2/2.3 dry and 10YR 3.3/2.3 wet. Albedo measurements were made during selected mornings with clear skies in 1995 and 1996. The mean albedos of reference smooth surfaces (<2 mm sieved soil) were 0.279 and 0.155 for dry and wet Gila soil and 0.221 and 0.114 for dry and wet Pima soil. Four tillage conditions were studied: rough plow, disk, disk–disk, and seedbed. Tillage direction was north–south in 1995 and east–west in 1996. The goodness of fit of linear relationships determined between mean albedo and surface roughness, measured with a roughness meter and reported as the root mean square deviation, were relatively high; however, the slopes for the regression equations were different for the two soils and the two moisture conditions. The different slopes indicate that the sensitivity of albedo to surface roughness was highest for the most reflective surface (dry Gila soil) and lowest for the least reflective surface (wet Pima soil). The albedos were on average 27, 18, 10, and 8% lower for dry and wet rough‐plow, disk, disk–disk, and seedbed treatments, respectively, as compared with the albedo of the reference smooth soil. These reduction percentages can be used as a general guide to estimate the albedos of tilled soils similar to the Gila and Pima soils studied here.

The albedo of earth surface features, such as soil, is an important component of models that define land‐surface meteorological processes. If land surfaces have no vegetative cover, soil properties determine the amount of solar radiation absorbed or reflected. We evaluated the influence of two soil properties, soil color and soil moisture, on soil albedo. Two soil moisture conditions were studied, air dry and wet, defined as the condition when the water films are absorbed by the soil and no water glistens on the soil surface. The albedos for 26 U.S. soils were measured with an Eppley pyranometer, which integrates radiant energy in wavelengths between 0.3 to 2.8 μm. Soil colors were measured with a Minolta Chroma Meter and spectral reflectance curves from 0.45 to 0.9 μm (measured in 0.1‐μm increments) were determined with a multispectral radiometer. All measurements were made on <2‐mm smooth soil surfaces, and the dry and wet data were combined for statistical analyses. Soil albedos were significantly correlated with Munsell soil color value , blue , green , red , near infrared (NIR), , and sum of the four bands ; however, the slopes and intercepts for these relationships were different. The 52 spectral curves yielded nine cluster groups, which mostly related to the Munsell soil color value and soil albedo soil characteristics. The 0.3‐ to 2.8‐μm albedos of smoothed soils can be accurately estimated using the regression relationship: soil albedo (0.3–2.8 μm) = 0.069 (color value) − 0.114. Using the regression equations presented here, spectral reflectance data in selected visible and NIR bands can also be used to predict albedo.

The spectral reflectance characteristics of 11 rainfall simulator plots were measured with a hand-held radiometer on semiarid rangeland surfaces, of the U.S. Department of Agriculture, Agricultural Research Service, Walnut Gulch Experimental Rangel- and Watershed, Tombstone, Arizona. Rainfall simulations were made in Map and June 1984 for plots characterized by natural vegetation, natural vegetation clipped and removed, and natural vegetation clipped and rock fragments >5 mm removed. In 1994, reflectance and cover data were collected on the same plots to evaluate changes over time. Measurements were taken at 38-40 degrees and 74-77 degrees sun elevation angles with a four-band hand-held radiometer (blue 0.45-0.52 mu m, green 0.52-0.60 mu m, led 0.63-0.69 mu m, and near-infrared 0.76-090 mu m). Correlation and regression relationships were computed between spectral reflectance and percent soil, rock, and vegetative cover; percent runoff; and eroded sediments. Highly significant correlations were measured between vegetative cover and percent runoff; relationships with soil-rock cover and eroded sediment were poorly correlated. The normalized difference vegetation index (NDVI) was the best predictor of percent vegetative cover, with shrubs-forbs being most strongly correlated to reflectance. The regression relationships between 1984 and 1994 spectral reflectance and vegetative cover were very different, even though cover percentages were similar. The amount of standing live and dead biomass and the proportion of green biomass strongly affect spectral reflectance, and the 1984 and 1994 conditions were quite different. Spectral reflectance data can be used to predict rangeland cover characteristics, which in turn, can be used to determine parameters needed for models that predict hydrologic processes on rangeland surfaces.

T-4464 buffelgrass (Cenchrus ciliaris L.), a perennial bunchgrass from Africa, has been extensively seeded throughout Mexico. After establishment and grazing, T-4464 either persists with time and actively invades surrounding areas (spreads), persists with time but does not increase (persists), or declines with time and all plants die (dies). To help land managers select high-potential seeding sites, we classified 139 seeding sites in three survival regimes: (i) spreads, (ii) persists, and (iii) dies. In previous research, we identified a relationship between plant survival and organic C. This research was designed to identify relationships between organic C and soil color. Single comparisons between organic C and Munsell hue, value, chroma, and reflectance in dry and moist soils were poor predictors of plant survival. To predict buffelgrass survival among the three survival regimes and between spreads and dies, we used discriminant function analyses. In dry soil, a model including value and chroma correctly classified 53% (Wilke's lambda = 0.8) of the seeding sites in the three survival regimes, while in moist soils, value and reflectance components correctly classified 61% (Wilke's lambda = 0.7) of the seeding sites. A dry soil model including value, chroma, and reflectance correctly classified 81% (Wilke's lambda = 0.7) of the seeding sites between spreads and dies, while a moist soil model, including the same components, correctly classified 83% (Wilke's lambda = 0.6) of the seeding sites. Survival regime selection with multiple soil color components prior to brush control and sowing will reduce adverse economic and environmental consequences and enhance long-term beef production.

The reflectance of radiant energy from the earth's surface in sparsely vegetated arid rangelands is determined by the characteristics of the soil and geologic material on the land's surface. This study measured the color characteristics of earth surface materials collected from a semiarid rangeland in southeastern Arizona and compared these colors to digital numbers recorded by Landsat. Other parameters including particle size, slope, and vegetation were also evaluated, but the color characteristics of the fine earth soil and rock fragments measured with a colorimeter were most strongly correlated to Landsat digital numbers. The numerical values of the color components (hue, value, and chroma) for three different soil and rock fragment size fractions were related in a multiple linear regression equation to Landsat digital reflectance numbers. The R(2) for Band 4 (0.5-0.6 mu m), Band 5 (0.6-0.7 mu m), Band 6 (0.7-0.8 mu m), Band 7 (0.8-1.1 mu m), and the sum of the four bands were 0.85, 0.69, 0.71, 0.68, and 0.75, respectively. The color of earth surface features in sparsely vegetated land areas should be precisely and accurately determined because of its very strong correlation with remotely sensed spectral data. The use of colorimeters to quantify the color of earth surface features will significantly help in evaluating remotely sensed data, particularly for landscapes in arid regions.

We prepared sets of <2 mm soil samples, distributed them to soil scientists, and asked them to determine the dry and moist Munsell color of each soil. We observed that soil scientists agreed on the same color chip for a single color component (hue, value, or chroma) 71% of the time, and there was an average of 52% agreement for all three color components. The standard deviation (SD) varied from 0.45 (value-moist) to 0.68 (chroma-moist) with an average SD of 0.57. Regression equations were computed that compared the mean soil color with the nearest color chip noted by an individual soil scientist, and the coefficient of simple determination (r 2 ) ranged from 0.49 (chroma-moist) to 0.79 (value-moist and dry). When “in-between” colors were estimated the r 2 improved and ranged from 0.70 (chroma-moist) to 0.95 (value-moist). A detailed evaluation was made of data from a commercial tristimulus colorimeter, and results were compared to colors described by soil scientists. The r 2 ranged from 0.88 (chroma-moist) to 0.96 (value-dry); however the slopes and intercepts were different. Commercial colorimeters have great potential as tools for measuring soil colors, but field colors by soil scientists are not identical to instrumental data. They differ because the sensor, light source, and angle of light refraction are different for each color measurement method. The inherent complexity of color identification by humans vs. instrumental measurements is also a contributing factor.

This paper evaluates the ability of college students to estimate soil texture. The students (115 total) were enrolled in the Soil Morphology, Classification, and Survey classes taught at the University of Arizona from 1983 to 1987. Each semester the students were provided in the laboratory with 30 known reference samples during the first 2 wk of the semester. At the end of the third week, they were asked to estimate the percent sand, silt, and clay in 15 different samples, and also to identify the correct soil textural class. During the last week of the semester in 1983 and 1984, estimations on an additional set of 15 samples were made by 54 students. Simple linear regression equations and correlation coefficients were calculated that compared the student estimations of the percent soil separates to laboratory analyses. This paper further compares student data to estimates made by professional field soil scientists. The correlation coefficients and y intercepts at 3 wk, the end of the semester, and for professional soil scientists were respectively: percent sand—0.85/−3.9, 0.85/0.47, and 0.83/−2.5; percent clay—0.74/2.6, 0.79/4.5, and 0.81/3.1; percent silt—0.63/6.5, 0.55/8.1, and 0.72/5.7. The mean percent correct (using the 12 textural triangle classes) was 39% at 3 wk, 45% at the end of the semester, and 46% for professional soil scientists. Results indicate that with 3 wk of practice and the availability of good reference samples, students can attain a level of proficiency comparable to the professional soil scientists. Suggestions are made as to how some of these data could be used in grading field exams and soil judging contest score cards. Similar data collected on students in the basic soils class are also included in the paper.

The American Society of Agronomy committee A335.4 was established and charged to evaluate and recommend ways to enhance the effectiveness of the Journal of Agronomic Education (JAE). The objective of this article is to present the surveys and survey results used to evaluate JAE. Of approximately 12 800 surveys printed and distributed in Agronomy News (AN), only 234 (about 2%) were returned. Respondents were asked to compare the importance of JAE with Agronomy Journal (AJ), Crop Science (CS) and Soil Science Society of America Journal (SSSAJ). The ratings used ranged from 1 (lowest) to 5 (highest), with AJ, CS, and SSSAJ being assigned a relative rating of 3. Surveys from AN indicated that the rating by JAE subscribers (2.50) was significantly higher than the rating for those who did not subscribe (1.91). A second set of surveys plus four complimentary copies of JAE were mailed to 125 departments. One survey was to be completed by the department head and a different survey by a top resident-instruction educator, an outstanding researcher, and an excellent extension educator. Department heads were asked to rate JAE relative to other ASA journals before and after perusing the complimentary copy. The 63 forms (50%) returned indicated that the rating for JAE increased from 2.32 in the preliminary rating (before perusing a complimentary copy) to 2.54 (P level = 0.094) for the rating after viewing the JAE issue. Major area of previous department head responsibility did not affect the responses. Faculty participants returned 170 out of 375 forms (45%). Reviewers involved primarily in research gave significantly lower ratings (2.05 versus 2.58), tended to consider JAE articles less important for tenure and promotion (51% vs. 84%), and subscribed at a lower frequency (8% vs. 66%) than those in teaching. The 10 recommendations presented by the committee were unanimously accepted by the ASA Board of Directors at the 1987 Annual Meetings.

AbstractPedon descriptions, vegetation transect information, and Landsat digital data were obtained for 110 sites on the Tonto National Forest in central Arizona. Using the field and satellite data, 33 variables were evaluated and prediction models were generated using stepwise multiple regression techniques. The following six factors explained 84% of the variability within the sum of the values for the four Landsat spectral bands: sum of brush and forest crown densities, elevation, surface color, rock type, cobbles on the surface of the site, and grass cover. Seven factors explained 81% of the variability for the ratio of Bands 4 plus 5 to Bands 6 plus 7: percent clay in the surface horizon, percent fragments > 2 mm in the surface horizon, the sum of forest and brush crown densities, pH of the surface horizon, color of the surface horizon, litter cover, and site aspect.