Permittivity Of Green Wood

The permittivity of green wood is an especially important factor affecting the microwave remote sensing of forests [1). Recently, EI-Rayes and Ulaby [2] described a semi-empirical model for evaluating the microwave dielectric behavior of vegetative material as a function of water content, frequency, and temperature. Using data acquired primarily from corn-leaf measurements, the model was so generalized as to suggest its validity for such forest components as green wood and foliage as well. However, a more recent study by Altman and Schneider comparing the El-Rayes/Ulaby model with an extensive data base of forest permittivity measurements suggests that still better agreement can be achieved if the EI-Rayes/Ulaby curves are lowered and the minima of the loss factor shifted lower in frequency. These modified results are consistent with the relatively smaller bound-water volume fractions of green wood as determined by Trapp and Pungs [3]. In an effort to view the wood more realistically, an earlier model based on Tinga's multiphase dielectric mixture theory [4,5) which shows excellent agreement with measured data at L-band for moisture contents below thirty percent is extended to green wood with gravimetric moisture contents up to as high as 200 percent (dry basis). The key to a good model for wood appears to be the use of proper values for the bound water and a proper apportioning of the bound and free water. Both EI-Rayes/Ulaby and Tinga give approaches to solving this problem. These methods and the reSUlting model data are compared with the experimental data base in the literature. Introduction CyberCom Model III for the permittivity of green wood and foliage e was based primarily upon Broadhurst's tulip-tree leaf and branch data [6]. According to that model, within the frequency band 200-2000 MHz the real and imaginary parts of the permittivity f (related to the susceptibility through the relation f X + 1) are given, respectively, by .' 40 (1) En _ (l.S/f ) + (2f )/[1 + (f /20)2) (2) GHz GHz �z University of Alberta Edmonton, Alberta CANADA T6G 2G7 Tel: (403)-432-2404 Fax: (403)-432-7219 The real part is clearly frequency independent; the imaginary part is nearly inversely proportional to frequency below 500 MHz where conduction losses dominate, but following a minimum near 800 MHz, increases nearly linearly with frequency up to 2000 MHz due to relaxation losses attributable to water (both bound and unbound). The model does not explicitly account for moisture content within the wood because there had been good reason to suppose that the moisture content of green wood showed little seasonal variation [7,8] . This, of course, is contrary to the large seasonal variation of moisture content exhibited by grasses and legumes (e.g., corn, soybeans, and wheat). However, conflicting data show fairly substantial seasonal variations in tree water content [9]. Further, the water content of green wood (and so the permittivity) and its spatial distribution within the tree is· species dependent; for example, the moisture content of softwoods (conifers) is usually greater than that of hardwoods (deciduous trees) as is their heartwood-to-sapwood water content ratio [refer to Figure 1). Both James [10], at low frequencies, and Tinga et al [4], at low moisture contents, offer water­ content-dependent models to fit their data, but neither provides parameter values. El-Rayes and Ulaby [2] consider high frequencies and moisture contents, with explicit, manageable equations for the permittivity as follows: where f f + V f + v. (3) v r fw t b b total permittivity non-dispersive residual component f free-water permittivity f , bound-water permittivity b v free-water volume-fraction r .. Vb bulk-vegetationfbound-water volume At room temperature (22°C) for free water having conductivity a (in siemens) • _ 4.9 + 75/(1 + a2) + j(75a/(l + a2) + ala) (4) f (5) e "III "" "! -, .� III � .; , c .c '" 2849 OJ IS) I J ..:r ",M H