energy exponential relationship presented by Nelson 

 (1983). The decrease in Gibbs free energy per gram of 

 sorbed water is described by (Nelson 1983, 1984; Skaar 

 1972): 



AG = - (/?r/M)ln(RH/100), cal/g (1) 

 where R is the universal gas constant, 1.9872 cal °K/mole; 

 Tis the absolute temperature, (°C + 273 = °K); M is the 

 molecular weight of water, 18.0153 g/mole; and RH is the 

 relative humidity in percent. Limits on AG were set by 

 considering the values as RH approaches zero and 

 100 percent. When the RH goes to zero, AG would become 

 negatively infinite but is limited to a value that is defined 

 when the RH has reached a small but finite value of 0.66 

 percent. Therefore AG^ represents the Gibbs free energy 

 change possible at very dry conditions: 



AG = AG^ when MC = 0. (2) 



AGq is determined from a plot of ln( AG) versus MC and 

 extrapolating to MC = 0. 



When the RH approaches 100 percent, the MC ap- 

 proaches fiber saturation moisture content, MC^, and the 

 Gibbs free energy goes toward zero so: 



AG = when MC = MC^. (3) 



Nelson (1983, 1984) found a simpler logarithmic model 

 to have better linearity with In(AGp) = (or AG = 0) when 

 MC equals MC,: 



In(AG) = ln(AG„) (1 - MC/MCp. (4) 



To be as accurate as possible, equation 4 was modified 

 by defining MC^ as the moisture content value when 

 In(AG) = and is only an approximation of fiber satura- 

 tion moisture content, MC, under conditions of desorption. 

 For adsorption, MC^ approximates the product of MC, and 

 the hysteresis ratio (the average ratio of adsorption to 

 desorption moisture contents). Equation 4 then becomes: 



In(AGo) = ln( AGq) ( 1 - (MC/MC^)) (5) 



and may be applied to adsorption and desorption EMC 

 data. 



This equation can be modified to a form suitable for 

 testing experimental data by least squares regression 

 analysis when A = InAG^) and B = - {A/MCJ: 



ln(AG)=A-^S(MC). (6) 



Values for AG were determined by equation 1 for each 

 temperature and relative humidity tested. The AG 

 values were combined with the observed EMC values 

 to develop the data sets used to determine the coefficients 

 A and B. Six to ten observations of EMC were used in 

 each combination (table 2). Analysis showed that MC 

 was a predictor of In(AG) for species, temperature, and 

 RH. Within the range of temperatures tested, I developed 

 a quadratic or linear equation to predict A and B for a 

 given, ''K, temperature. The form of the equations for 

 predicting A or S is: 



A = C^ + C^^j(TEMP) + C„^2 (TEMP^) (7) 



where 



TEMP = fuel surface temperature, °K 



C^, C^^j, and C^^^ = constants of regression. 



With a given temperature and RH, we could then deter- 

 mine In(AG), In(AGp) =A and MC^ = - (A/B) and estimate 

 MC using: 



MC = MC^ (1 - (ln(AG)/ln(AGo)). (8) 



This provided the means to evaluate the variability in 

 EMC of foliar litter fuels, check for similar groups of lit- 

 ter, define how both the sorption process and the state of 

 weathering influence EMC, and show how the EMC's of 

 foliar litter relate to the wood fine fuel EMC used in the 

 NFDRS. The EMC's of the forest foliage litter fuels are 

 presented in the accompanying figures and referenced to 

 the EMC of wood at the same temperature. Curves for 

 wood were generated by regression equations produced 

 in the same way as for foliar litter but using the EMC 

 data in table 3-4 of the Wood Handbook (USDA FS 1974). 



RESULTS 



All the tests showed a typical sigmoid-shaped curve for 

 EMC as a function of RH when EMC is assumed to ap- 

 proach zero as RH nears zero (fig. 2). Nearly all the foliar 

 litter samples showed higher EMC's than calculated by 

 the method used for wood sticks in the NFDRS. Recently 

 cast Douglas-fir needles showed a sorption change in 

 EMC at 300 °K (80 °F) very similar to that of wood used 

 in the NFDRS (fig. 2). At the same conditions, western 

 larch needles showed the highest EMC's (fig. 2). The 

 other litter samples tested have data located between 

 these two and show that the EMC's can differ by 2 to 6 

 percent, depending on the litter samples. The weathered 

 samples showed similar responses, but are shifted toward 

 higher EMC's and have less hysteresis among the foliar 

 litter samples (fig. 3). The adsorption and desorption 

 curves in figures 2 and 3 show the hysteresis loop to be 

 2 percent or less EMC for RH from 10 to 90 percent and 

 a temperature of 300 °K (80 °F). 



The EMC's at 278 °K (40 °F) and 300 °K (80 °F) were 

 higher than those of wood, but at 322 °K (120 °F) a signifi- 

 cant decrease in EMC occurred (fig. 4). This decrease is 

 more pronounced than reported for wood in the Wood 

 Handbook (USDA FS 1974) and as used in the NFDRS. 

 At 278 °K (40 °F), moisture contents of the foliar litter 

 samples could be 4 percent higher at low RH and as much 

 as 7 percent higher at high RH. As the temperature in- 

 creases, the EMC decreases until at 322 °K (120 °F) all 

 four litter samples have EMC's lower than wood at RH's 

 above 40 percent. Cheatgrass had a small change in 

 EMC, 4.2 to 10.6 percent, as RH went fi-om 12.7 to 88 

 percent at 322 °K (120 °F) (fig. 4). Ponderosa pine 

 needles' EMC predicted values from equations 6, 7, and 

 8 show the decrease in EMC at 10, 30, and 70 percent 

 RH as the temperature increases (fig. 5). 



3 



