eters, defects, etc. may be determined by 

 subsampling a portion of the sample line. 



Line-intersect sampling is highly recom- 

 mended for several reasons : 



1. For the same level of accuracy in volume 

 estimation, the line-intersect method re- 

 quires only one-fourth to one-half the 

 time needed for fixed-area plots. 



2. There is no problem with pieces overlap- 

 ping a boundary as there is with fixed- 

 area plots. 



3. When the line-intersect method is used, 

 there is no difficulty in determining what 

 has or has not been tallied, because meas- 

 uring proceeds systematically along the 

 sample line. 



However, less attention has focused on a 

 general system for compiling a useful picture 

 of residue conditions for a particular area. 

 Bailey does describe some of the techniques 

 needed in such a system, but all the details 

 are not combined in a step-by-step manner. 



The REST System 



The REST (Residue ESTimation) system 

 described in this note is one approach for 

 putting residue information into a format 

 that will be useful to industry, resource ana- 

 lysts, and resource planners. REST incor- 

 porates the methods developed by Bailey, 

 along with some additional measures for 

 assigning the residue volumes (or weights) to 

 the appropriate diameter and length classes. 

 This information can be subdivided further 

 by species or species groups and by type 

 (sawable or chippable only, straight or 

 crooked, etc.) . 



Although this sounds like a straightfor- 

 ward procedure, the computations are lengthy 

 because of the assumptions on which the 

 line-intersect method is based. By using the 

 line-intersect sampling scheme, we are saying 

 that a satisfactory volume (or weight) esti- 

 mate is possible when only the diameter at 

 point of intersection is measured. When a 

 total per-acre figure is our goal, we have no 

 problem. However, when we subsample the 

 residue and attempt to group the data by 

 length class, we run into difficulty because 



the probability of intersecting a piece is pro- 

 portional to its length. 



Bailey resolved this by what he calls a five- 

 step rectification procedure. The computa- 

 tions involved are incorporated into REST 

 along with provisions for assigning the 

 pieces to the proper diameter classes. REST 

 is primarily the result of taking what has been 

 developed by others and organizing a system- 

 atic means of translating field data into a 

 workable description of residue conditions. 



The steps in the REST system have been 

 programmed in FORTRAN IV for use with 

 an IBM System/370 computer. However, the 

 aim in programming was to achieve general- 

 ity so that REST could be used with most 

 computer facilities (fewer than 45,000 bytes 

 of core storage are needed). Flexibility was 

 another programming goal, so that REST 

 should fill the need of users in any timber- 

 growing region. 



Features of REST 



Virtually any approach to field measure- 

 ments with the line-intersect method is com- 

 patible with REST. Sample lines (sampling 

 units) may be referenced to a grid system 

 and oriented unidirectionally, randomly, or 

 paired at right angles to each other. Another 

 possibility is to use a continuous line with a 

 random change in direction after a predeter- 

 mined number of feet or chains have been 

 measured. The REST system is set up so 

 that information collected from a subsample 

 of the sampling units can be used easily for 

 developing a conversion ratio — to adjust 

 cross-section area at the point of intersection 

 to correspond with the mean of piece end 

 areas and to adjust for defective wood (Bailey 

 1970b) — and for developing length and diam- 

 eter distributions etc. 



The user will have to decide on appropriate 

 procedures. However, REST is designed to 

 handle the summarization for any field lay- 

 out and for any intensity of subsampling 

 (from to 100 percent). Other specific fea- 

 tures are: 



1. Use of the conversion ratio (as recom- 

 mended by Bailey) is optional. 



2 



