->» 



Appendix 



<&r 



The Field Survey and Its Accuracy 

 Forest Land Area and Timber Volume 



THE ESTIMATES of forest acreage and timber 

 volume are based upon a systematic sampling 

 system involving a forest-nonforest classification 

 on aerial photographs and on-the-ground measure- 

 ment of quarter-acre sample plots. The sample plots 

 were taken in pairs at and near the intersections of a 

 grid of east-west and north-south lines spaced 3 miles 

 apart. Field work in the State was begun in July 

 1948 and completed in October 1950. 



Accuracy of the estimates may be affected by two 

 tvpes of errors. The first type stems from the use of a 

 sample to estimate the whole and from variability of 

 the item being sampled. This type is termed sampling 

 error; it is susceptible to a mathematical evaluation 

 of the probability of error. The second type of error 

 derives from human mistakes in measurement, judg- 

 ment, arithmetic, or recording, and from limitations 

 of method or equipment. Effects of this second type 

 of error — often referred to as reporting and estimating 

 error — are not susceptible to a mathematical ap- 

 praisal, but the Forest Survey constantly attempts to 

 hold such errors to a minimum by proper training and 

 good supervision, and by emphasis on careful work. 



Statistical analysis of the data, using random- 

 sampling formulas, indicates a sampling error of 0.4 

 percent for the State estimate of total forest area, 1.7 

 percent for total cubic-foot volume, and 2.3 percent 

 for total board-foot volume. However, because a 

 systematic sample is generally more efficient than a 

 random sample of the same size, these estimates of 

 sampling error may be considered as setting an upper 

 limit of error, rather than as expressing the actual 

 probability of error. 



As the State total acreage and volumes are broken 

 down by Survey region, county, forest type, species, 

 and other subdivisions of the data, the possibility of 

 error increases and is greatest for the smallest items. 

 The order of this increase is suggested in table 7. 



Table 7. — Maximum sampling error to which Survey estimates c 

 Tennessee's forest area, cubic-foot volume, and board-foot volume mc 

 be liable l 



Commercial forest 



Growing stock 



Sawlog growing stock 



Area 



Sampling 

 error 



Total 



Sampling 

 error 



Total 



Sampling 

 error 



Thousand 

 acres 



Percent 



Million 

 cu. ft. 



Percent 



Million 

 bd.-ft. 



Percent 



12,500 



0.4 



6,000 



1.7 



16,000 



2.3 



5,000 



.6 



3,000 



2.3 



10,000 



2.9 



2,000 



1.0 



1,000 



4.1 



3,000 



5..4 



500 



2.0 



400 



6.4 



1,000 



9.5 



50 



6.4 



40 



20.3 



100 



29.3 



1 Liable on a probability of 2 chances out of 3. 



Growth 



Gross growth estimates are based on radial-growth 

 measurements of mechanically sampled trees on 

 Forest Survey plots. Measured growth over the past 

 10 years was used for projecting growth 10 years 

 ahead. The difference between present and projected 

 volume of the sample trees was reduced to an average 

 annual percentage increase and applied directlv to 

 the current inventory volume. Net growth was cal- 

 culated by subtracting from gross growth an allowance 

 for mortality based on the plot inventory of trees that 

 had died in the previous 4 years. No attempt was 

 made to calculate sampling error in the growth 

 estimates. 



Production and Annual Cut 



Annual cut estimates are conversions of production 

 estimates. Sawlog production in 1949 was estimated 

 from a survey of lumber production conducted by the 

 U. S. Bureau of the Census. For other commodities a 1 

 canvass of 1949 timber production was conducted bv I 

 the Forest Service, with some assistance from the I 

 Division of Forestry Relations of the Tennessee Valley 

 Authority. Commercial log and bolt production, I 

 other than for lumber, was obtained by a 100-percent 1 

 canvass of establishments or producers. Production 



40 



Forest Resource Report No. 9, U. S. Department of Agriculture 



