Th e b a c k g r o u n d a n d p u r p o s e o f t h e i n v e s t i g a t i o n s
This report is the third part of the results from the Pr o j e c t “ Fo r e s t a n d Fo l k” . A research project aimed at producing a better basis for deci
sions in the field of forest recreation. The project has been initiated and financed by the Da n i s h Ag r i c u l t u r a l a n d Ve t e r i n a r y Re s e a r c h Co u n c i l,
and since 1980 financed by grants from the De p a r t m e n t o f Fo r e s t r y at the
Ro y a l Ve t e r i n a r y a n d Ag r i c u l t u r a l Un i v e r s i t y in Copenhagen, the
Mi n i s t r y o f Ag r i c u l t u r e, and the Mi n i s t r y o f t h e En v i r o n m e n t.
The literature studies, form ing the basis for the investigations, have beén published as separate working papers in Danish only (Koch 1975, 1976a, 1976b, and 1977a). In part I (Koch 1978) a description was given of the Danish population’s use of the forests for recreation. Results from that survey and results from the present part I I I were used in part I I (Koch 1980), in which the geographical variation in the intensity and charac
teristics of forest recreational use was described for each region of Denmark.
In that way the overview of how the various investigations have been inte
grated is given in English in part I I (cf. Koch 1980, p. 203— 206) and in a summary article (Koch 1983/84) (see also Christensen 1983/84). An over
view article has also been published in German (Koch 1982b).
Part I II finishes the investigations into the use of the Danish forests for recreation carried out at the Project “ Forest and Folk” . The purpose of part I I I is to document the results of data collections at (1) the project’s five counting stations, and (2) some m anual recordings in five selected forest areas. The aims of each of the sub-investigations are given in the attached translation of figure I I .
In forthcoming publications from the Project “ Forest and Folk” it is planned to discuss preferences/ perceptions and, based on the data presented in part I, II, and 111: recreational use models and economic analyses.
Researchers who should fin d our data suitable for the last two study pur
poses are invited to participate in that work.
Me t h o d s o f t h e v a r i o u s i n v e s t i g a t i o n s
The following explanatory and elaborating notes are given to the sche
matic outline of the investigation methods presented in the translation of figure I I :
LOCALITY
1) Counting station No. 0 is an ordinary automatic traffic counter with one inductive-loop (magnetic-loop) detector. The counting station is similar to the stations normally used by the Danish road authorities who also since A pril 1, 1976 have operated and maintained counting station No. 0. It records hourly on paper the number of vehicles passing, but only the number of car visits per day has been used in the analyses. The mean counting error is estimated at — 3 %.
2) Counting stations Nos. 1, 2, 3, and 4 operate according to the n e t c o u n t p r o c e d u r e proposed by W agar & Thalheimer (1969). This counting principle is illustrated in figure 3 (see also figures 8, 12, 25, and 26), and it is essential to note that not only the number of car visits is recorded, but also the number of car visitor hours and the mean length of stay per car visit (for definitions of the units see e.g.
Bureau of Outdoor Recreation 1975, Koch 1978, p. 296 and 353, and 1980, p. 269). The mean counting error in counting the number of car visits variates for the four counting stations from — 0.3 % to 0.5 % (cf. table A.1.1). W hen estimating the number of cars present
— and hence the number of car visitor hours — it is necessary to correct for these counting errors, cf. figure 8 (and tables A.2.1—
A.2.2). This has been done by smoothing, and the statistical model used for smoothing for counting stations Nos. 1, 2, and 4 has been developed and documented in English by Andersen & Rudemo (1982).
For counting station No. 3, the absolute counting error is so large (cf. table A.1.1) that the model is not usable for the counts from this station. Instead a much more primitive method of smoothing has been used for counting station No. 3 (see also appendix A.7). The mean error in registering the number of cars present is estimated at approx. — 1.5 % for counting stations Nos. 1, 2, and 4, and approx.
— I I % for counting station No. 3 (cf. table A.2.3 and figure A.2.1).
3) The distribution of the missing observations at the counting stations is illustrated and documented in figure 9 and tables A.3.1 and A.3.2.
The correction for the missing observations has been carried out in two different ways: (1) By a weighting with the reciprocal value of the percentages given in tables A.3.1 and A.3.2 (for the results pre
sented in figure 15 and in section 7.2— 7.5), and (2) by an estimation of the missing values (for the results presented in figures 16, 17, 19, 21, and 22 and in section 7.6). Rather simple m ultiple linear regres
sion models have been used in the estimation of the missing values, cf. tables A.3.3— A.3.5 and figures 10, and A.3.1— A.3A. These models could be refined, cf. appendix A.7.
4) The manual recordings in the forest area “ Ärhusskovene” were carried out on not randomly selected registration-days (cf. table
AAA ), and as interval-counts in randomly sampled forest compart
ments (cf. figure 55), which makes the counting unit poorly defined (the probability of getting interviewed increases with the forest area covered in a given visit). Also the interviewing form used in this investigation has many shortcomings (cf. appendix A.7 and en
closure 1.2), and as a whole the investigation in “ Århusskovene”
should be considered as a preliminary investigation only.
5) The m anual recordings in the other forest areas have been collected by means of the same method: The registration-periods and the registration-places have been selected by stratified random sampling (cf. tables A A.2— AAA). For the forest area “ Gammelkjøgegaard”, the investigation covers the period 1/5— 30/9 1982 only, and not one year as is the case for the other forest areas. As registration-places all exits from the forest areas have been used (cf. table AAA ). Each m anual counting has taken 15 minutes and has resulted in a record
ing of every person leaving the forest area in the sampled registration- period at the sampled registration-place. The recording form appears in enclosure 1.1. At each counting one person was sampled for inter
viewing among everybody who was judged older than 15 years in the first registered group. No more than one interview was carried out per counting, even in the case of a non-response interview. The inter
viewing form, which has been slightly modified from one forest area to the other, appears in enclosure 1.3. Between each counting 15 minutes were set aside for transportation to the next registration- place (cf. the notes at table AAA).
6) The specification of the non-response at the interviewing is given in table A.5.1 (see also table B.14-). The response-percentages are acceptable, and a correction for the non-response is not considered necessary.
7) The registrations at the m anual recordings have been generalized in two different ways: (1) The interviews and the observations are weighted in accordance with the stratification in the sample design.
However, the sample design is complex and with some shortcomings, and adjustments made for practical reasons (see above). This makes the calculation of confidence intervals for the estimates an impos
sible task (cf. appendix A.7). Nevertheless, the results are considered acceptable for the comparisons which have been carried out (cf.
appendix B.2). (2) The manual countings have been generalized by estimation of multiple linear regressions with the recordings at counting stations Nos. 1, 2, 3, and 4 of the number of cars leaving in the same 15-minute periods (cf. table B.13 and p. 166). This is not possible for the forest area “ Århusskovene”, due to the poorly
defined counting unit in this investigation (cf. point 4 above). The standard deviation of the number of visits per year is also given in table B.13. However, it should bé noted that the risk of non-sampling errors is probably considerable for these results ( e.g. in the estima
tion of the so-called reduction factor, cf. table B.13).
Re s u l t s
Results for the counting stations are presented in figures 12— 52 and tables B.1— B.12, while results of the m anual recordings in selected forest areas are given in figures 53— 57 and tables B.13— B.21. The control of the estimates in part I I of the yearly number of forest visits and visitor hours is mostly mentioned in the Danish text only ( cf. p. 246 f and also tables B.13, B.15, and B.16). However, the m ain conclusion is that the coincidence be
tween the results of the widely different investigation methods is acceptable.
The deviations for the few re-investigated forest areas are smaller than would be expected from the assumed error of the total number of visitor hours in part II (cf. p. 246 f and Koch 1980, p. 206).
Some comparisons with other selected investigations have been carried out. These, too, are mostly discussed in the Danish text only, and in tables C.2— C.3. Data about the other investigations is summarized in table CA.
A ll the secondary enclosures (cf. p. 128) — containing more elaborate methodological descriptions, the computer programs, and more detailed results — are published separatelg. They can be obtained from the project.
Pr o j e c t Fo r e s t a n d Fo l k
The Danish Forest Experiment Station Springforbivej 4, DK-2930 Klampenborg.
Det forstlige Forsøgsvæsen. X X X IX . H. 2. 31. marts 1984.
10. EFT ERSK RIFT