A primary objective of this study has been to rate selected photo- stratification schemes 

 as simply and as directly as possible. Although complete cost data were not obtained on this 

 study, no discussion of stratification methods would be complete without some reference to costs. 



RELATIVE COSTS 



In mountainous regions , field plot costs depend more on travel time than on time actually 

 spent in plot measurement. Based on the St. Regis study, total field measurement time can be 

 expected to average about 14 man-hours per location. The Forest Survey records for this area 

 indicate average costs of $70 per location with some locations exceeding $100. 



Photo interpretation costs are affected little by topography, but depend almost entirely on 

 the kind of measurements and classifications made. In commercial forest stands where cubic - 

 and board-foot volumes are estimated by means of precise photo measurements and these esti- 

 mates and measurements are used to stratify the plots , studies (7) show that experienced photo 

 interpreters can average about five plots per man-hour. Where classifications are made with- 

 out measurement, or in areas where many plots are nonforest, much higher averages are 

 possible and sustained rates for stereo classification may reach 200 plots or more per man-day. 

 The skill and precision expected of a good photo interpreter may require a year or more of in- 

 tensive training and experience, but this interpreter should then be able to measure and classify 

 plots for about 70 cents each or 1 percent of the cost of fieldwork. 



The cost of aerial photos can generally be ignored in any evaluation of survey methods . 

 Aerial photos are working tools of the forester. They are obtained for many purposes in addi- 

 tion to forest inventory, and are readily available on most forests. In this particular case, the 

 photos used were originally purchased for field plot location in the St. Regis field study. Since 

 all photo methods were based on one set of photo measurements and classifications, and were 

 tested by field data taken on the same plots, the cost of aerial photos can have no significance 

 in these results. The same may be said for such routine chores as trimming and filing photos, 

 determining photo scale, and marking plots on both photos and maps. They are needed equally 

 in every method used. For example, no qualified photo interpreter would suggest plot 

 classification by any scheme in mountainous areas without knowing photo scale. 



Both photo and field plot measurements used in this study were made by personnel with 

 pay scales considerably above average. A comparison of costs is therefore more meaningful 

 when based on man-hours. Photo interpretation including stand heights measured by parallax 

 wedge, crown cover by comparison, and again by dot count, slope percent computed from hori- 

 zontal and vertical photo measurements, cubic- and board-foot volumes read from aerial tables, 

 and finally, classification of all plots by six different methods was performed at a rate of about 

 30 plots per man-day. 



As this study indicates, merely sorting these measured plots by different methods can 

 result in widely different field plot requirements for reliable volume estimates and correspond- 

 ing savings in overall costs. These costs can be compared in a simulated problem. Assume 

 100 unstratified field plots are required to obtain a reliable volume estimate and that measuring 

 the plots on the ground would require 1 ,400 man-hours of field time. To stratify these plots by 

 stand size and density using photo interpretation without measurement would cost 4 man-hours. 

 The gain is similar to that shown for map stand size and density (table 2) and should reduce the 

 needed field plots to 76 and the field cost to 1,064 man-hours. Measuring the plots on aerial 

 photos and then stratifying them into stand size and density would cost 27 man-hours. However, 

 when these strata are sampled in the field, the required plots could be reduced to 58 and the 



13 



