Doctoral Dissertations

Date of Award

6-1972

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Biosystems Engineering

Major Professor

John I. Sewell

Committee Members

John I. McDow, W. L. Parks, and B. A. Tschantz

Abstract

The introduction of remote sensing techniques as an analyst's tool may well revolutionize the world of environmental analysis. However, for such techniques to be applicable to each specific analysis, a valid cause-and-effect relationship or theoretical basis must be developed. For the techniques to become operational and reliable, these basic relationships must be defined, understood and properly calibrated. The application of remote sensing techniques to the detection and classification of soil moisture levels on fallow soils requires an understanding of the relationships that exist between soil moisture levels and electromagnetic energy that is reflected, absorbed and emitted. These relationships involve the microclimate of the soil and the soil and water system. Proper implementation of these relationships requires an understanding of the operation of sensors or energy detectors. This study was designed to give a theoretical treatment of the cause-and-effect relationships between the physics of the soil and the resulting energy responses to be expected in the visible, near infrared and thermal infrared wavelengths. Laboratory and field studies were performed to determine the feasibility of and the extent to which soil moisture levels of a fallow fine sandy loam Sequatchie soil could be detected and categorized using visible and near infrared photographic films and near and thermal infrared energy detectors (electronic scanner) as sensors. Color infrared film (Kodak type 8443) was exposed through four combinations of camera filters including Wratten No. 15 and BOB and Corning CS-1-59(3966). Other films used included infrared black-and- white (Kodak type 5424, exposed through Wratten filter No. 25) and regular color film (Kodak type 2448). Scanner detectors included the 2000- to 5000-nanometer range and the 8000- to 14000-nanometer range. Results of the laboratory study to determine the relationship between the percent of energy reflected as a function of soil moisture levels and wavelengths (visible and near infrared) indicated that the reflectance decreased with increasing soil moisture to approximately 20 to 25 percent, where it began to increase in quadratic fashion. The function describing the soil reflectance-soil moisture relationship at 575 nanometers wavelength is typical of several developed, and it may be written as R = 2.116 - 0.112(SM) + 0.00256(SM)2 where R is percent reflectance and SM is soil moisture in percent dry weight. It was concluded that reflectance measurements alone could not adequately predict soil moisture levels throughout the entire range from oven-dry to field capacity. Results of the field study showed that for a limited range of moisture levels (approximately 1 to 24 percent dry weight) , the response data could adequately predict the moisture level with or without thermal response data. A typical regression analysis using color infrared film (with filter combinations 15 plus CS-1 and 15 plus 80B plus CS-1), black-and-white infrared film and both electronic scanner detectors resulted in the following prediction equation for soil moisture at the soil surface (SMCl): SMCl = 2.12 - 0.310(GR58C) + 0.0115(BL15C)2 + 0.000663(RD58C)2 where all independent variables were color infrared film responses defined as GR58C = camera filter combination 15 plus 80B plus CS-1 analyzed using the green densitometer filter (representing red wavelengths), BL15C = camera filter combinations 15 plus CS-1 analyzed using the blue densitometer filter (representing green wavelengths) and RD58C = camera filter combination 15 plus 80B plus CS-1 analyzed using the red densitometer (representing the infrared wavelengths). Incoming solar radiation and radiometric temperature of the soil were also found to be important variables in predicting soil moisture using response data from remote sensors.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS