CONCLUSIONS 



Models developed for predicting water retention at 1/3 and 15 bars tension demon- 

 •trite the strong influence of organic matter on water retention in coarse-textured 

 ^ Lis the Idaho Batholith. The models were developed using data from disturbed 

 ^■^il samples, a problem that could not be overcome due to lack of cohesive strength of 

 ^[^e soils. At 1/3 bar, the adverse effects of the disturbed soils on retention measure- 

 ments should be somewhat tempered by the weak structure inherent in the field soils. 

 '^f- 15 bars, almost all water is retained by adsorption, negating the problem of soil 

 ^'.imple disturbance. 



The original models were developed with the constraint that they fit theoretical 

 expectations over the ranges sampled for the three independent variables: percent by 

 weight organic matter, percent by weight fines (<50pm diameter), and bulk density. 



Two new data sets with wide geographic variability within the Idaho Batholith were 

 used to test the applicability of the original models. Both sets confirmed the general 

 curve form, and the original model performed reasonably well after least squares adjust- 

 ment. Tlie second new data set included three soils formed from loess having bulk densi- 

 ties and textures at the limits of the original data used in developing the models. 

 Predicted retention values at both energy values were surprisingly accurate for all 

 three loessal soils. 



For several reasons, the authors urge caution in the use of the models for predict- 

 in? hydrologic and ecologic factors such as water yield and "available soil moisture." 

 i:xperience in the Batholith has shown that the bedrock is generally highly fractured 

 :ind often highly weathered to great depths. Deep water storage and utilization of this 

 .>ater by deep rooted plants is universal throughout the Batholith. Drill cores recov- 

 ered from depths of 800 feet have shown hydrolysis of biotites and iron oxide stains 

 coating mineral grains (unpublished data on file, Intermountain Station, Boise, Idaho). 

 Secondly, the effects of disturbance on the soils disrupted the natural pore size dis- 

 tribution due to soil structure, which almost certainly affected the values determined 

 at 1/3 bar tension. Because of the coarse nature and inherent single grain structure 

 of the Batholith soils, this effect may be minimal, especially at low organic matter 

 and low fines percentages. 



The models do have value in estimating soil water regimes necessary in classifying 

 soils at the family level. In addition, they give some estimate of soil water retention 

 '.n the soil at two energy levels, and have practical value in estimating moisture 

 regimes for revegetation efforts. 



Because the regression models were interaction-sensitive, they allowed exhaustion 

 '^f information from the independent variables within the constraints of theoretical 

 ■-xpectation . Modest curve form bias and the close proximity of the standard errors 

 '^f estimate for the interaction models fitted to the original and to two new data sets, 

 tend to substantiate the general applicability of the original interaction forms. It 

 •■^ould, however, probably be advisable to regard the original model forms developed here 

 applicable on an interim basis, to be improved upon as more data become available. 



13 



