The current paper presents revised parameter estimates for the 
CACTOS growth model based upon a remeasurement of the permanent growth 
plots maintained by the Northern California Forest Yield Cooperative. 
The locations of the study plots are shown in Figure 2.

Figure 2.  Distribution of sample data for growth models: (a) stem 
analysis plots and (b) permanent plots.  Numbered to show townships and 
range coordinates of plot locations.  (Wensel, Meerschaert, and Biging, 
1987)
	
	Two distinctly different data sets were used to obtain the 
previous parameter estimates.  These are described by Wensel, Meerschaert, 
and Biging (1987) and are referred to here as STEM, the stem analysis 
plots and PERM, the system of permanent plots.  The initial measurement of 
the permanent plots is designated PERM0 and the 5-year remeasurement is 
designated PERM1.
	
	The STEM data base consists of tree data from 39 cluster plots from 
industrial forest lands.  The initial PERM data base, PERM0,  consists of 
measurements on 29,323 trees from 710 permanent plots.  Sample trees were 
selected from plots on lands of the industry members of the Northern 
California Forest Yield Cooperative (Figure 2).  The remeasurement data set, 
PERM1, consists of 569 plots with 19,363 trees.  The reduction in the number 
of plots from PERM0 to PERM1 resulted from not remeasuring the plot (75 plots, 
some of which were clear cut or were lost due to change in ownership), plots 
with missing data (52 plots), and plots that were damaged (14 plots).      
	
	Four regions of northern California were recognized for the purposes 
of this analysis.  Progressing counterclockwise around the northern end of 
California's central valley agricultural lands from due east of San Francisco 
to north of San Francisco the regions are as follows:  (1) the Sierra east of 
San Francisco, (2) the southern Cascades in the northeast, (3) the Shasta-
Trinity in the northwest, and (4) the Mendocino, north of San Francisco.  
Highways 32 and 36 from Chico to Susanville separate the Sierra and Southern 
Cascade regions, Interstate 5 separates the Southern Cascade regions and 
Shasta-Trinity regions, and highway 299 separates the Shasta-Trinity and 
Mendocino regions1.  The PERM data not only has more plots than the STEM data 
but it also has a wider distribution, particularly in regions 2 and 3.  

	The STEM data provided estimates of the height growth coefficients as 
well as tree volume and taper, crown geometry, and site index coefficients 
estimates.  The PERM0 data set, divided into two random subsets PERM0a and 
PERM0b, provided increment data for estimates of the tree diameter squared 
growth rates for the 6 conifer species.  The data sets were split to provide 
a sample for model formulation and model fitting and an independent sample 
for testing.  The STEM data were also used for initial estimates of the 
diameter squared growth rates but, representing a narrower range of tree sizes 
and conditions, these are likely to be less accurate when applied to the PERM0 
data.  Thus they were abandoned in favor of the PERM0 coefficients.  

	The remeasurement of the permanent plots provided the PERM1 data set 
which can be used to further develop the diameter and height growth parameter 
estimates.   As with the initial measurement PERM0, PERM1 was divided into two 
subsets, putting the odd-numbered plots into PERM1a and the even-numbered plots 
into PERM1b.   The number of trees in each of the data sets is given in Table 2.

Table 21.  Sample sizes for initial growth parameter estimates for tree height 
and DBH2 growth using the stem analysis (STEM) and permanent plot data bases 
(PERM0a and PERM0b, for the initial measurement and PERM1a and PERM1b for the 
remeasurement, where the first subscript indicates the measurement sequence and 
the second subscript denotes the first and second halves of the data set, 
respectively).
  		 	            Species	
		data set	PP	SP	IC	DF	WF	RF
	Height growth
		STEM	 	151	47	71	145	279	37
	 	PERM1a 		1261	347	820	585	1464	271
	 	PERM1b 		1178	373	841	791	1748	  (	
	DBH2 growth	
		STEM	  	532	100	420	386	914	126	
	 	PERM0a  	2064	905	1138	1465	3123	579
	 	PERM0b  	2139	  (	1176	1498	3166	  (	
	 	PERM1a  	1261	743	820	215*	1464	271
	 	PERM1b  	1178	  (	839	791	1748	  (

*  Only the potential was fitted as the competition components from PERM0 
   showed no significant difference.  Thus this number represents the number 
   of trees with little or no competition.
(  There were too few observations to split the SP and RF data sets to provide 
   independent test data sets for those species.