6. STUDY METHODOLOGY AND ANALYSIS OF DATA

* Plot/01: Wind-blown tree Plot/02: Tree knocked down by felling of neighbouring tree Note: In this study, workplace time excludes meal time because of its high variability.

On average, the workplace time per tree harvested was 13.90 minutes on Plot/01 (conventional logging) and 10.93 minutes on Plot/02 (planned harvesting). The time required per tree harvested serves as the basis for estimating production rates, treated in Section 6.2, and is distinguished from the time required to fell a single tree, shown in Table 7 for both plots.

In order to compare the time distribution of work elements and the time required for the felling of a single tree in both systems, the workplace time per tree rejected and the workplace time for bucking trees that were not felled (see the table above) have been excluded from calculations. The average time required to fell a single tree, and the distribution of the related time elements, are shown in Table 7.

Table 7. Time study results for tree felling under conventional logging and planned harvesting.
Work elements (classification)		Conventional logging, Plot/01 (without tree location map)		Planned harvesting, Plot/02 (with tree location map)
		[min]	%	[min]	%
Locating tree	(CW)	1.50	17.27	1.34	16.94
Preparing for felling	(CW)	1.07	12.30	1.33	16.76
Felling	(MW)	1.86	21.61	1.45	18.25
Bucking and delimbing	(MW)	1.84	21.22	1.76	22.28
Productive work time	(PW)	6.27	72.40	5.88	74.23
Searching for tree	(PL)	0.84	9.65		-
Evaluating tree	(OP)	0.05	0.60	0.08	0.97
Repairing equipment	(RT)		-		-
Maintaining equipment	(MT)	0.71	8.15	1.20	15.22
Refuelling chainsaw	(RF)	0.25	2.89	0.36	4.50
Clearing the site	(AW)	0.30	3.42	0.40	5.08
Cutting chainsaw free	(AW)	0.11	1.30		-
Other	(AW)	0.14	1.59		-
Supportive work time	(SW)	2.40	27.60	2.04	25.77
Time required to fell a single tree		8.67	100.00	7.92	100.00

The work element "searching for tree" refers to the time spent by the chainsaw operator searching for harvestable trees in the area. In accordance with the forest work study nomenclature (IUFRO 1995), this is considered supportive work since it does not directly contribute to the work progress on an individual work object and is to be distinguished from the work element "locating tree". The latter is part of the productive work time and of each individual cutting cycle where the operator moves to the next tree together with the chainsaw.

The productive time spent in felling a single tree on both plots is quite similar as expected due to similar stand conditions and the same felling crew involved. The major difference is that under conventional logging the chainsaw operator must spend an average of 0.84 minutes per tree searching for the next tree to be felled. When this non-productive time is included, the time required to fell a single tree then amounts to 7.11 minutes in conventional logging versus 5.88 minutes in planned harvesting.

A comparison of the time distribution of work elements that occurred in felling a single tree under the two logging systems reveals the following:

· the average time spent to fell a single tree under conventional logging is higher than in planned harvesting since a considerable share of the felling time is spent searching for harvestable trees;

· improper cutting techniques, limited skills in chain sharpening, and poor maintenance of the chainsaw result in a very high time share, namely 8-15%, spent on "maintenance" in this study as compared to only 1-3% found by Winkler (1997) in a similar study in the in the Brazilian Amazon.

6.1.2 Work and time study on timber extraction

The classifications and percentages of workplace time, excluding meal time, observed for each extraction work element on both the conventional logging operation and on the planned harvesting operation can be found in Appendix 3. Summary information on the time and production for the two plots is as follows:

Item	Conventional logging - Plot/01 (without tree location map)	Planned harvesting - Plot/02 (with tree location map)
Work time	13 h 15 min	17 h 21 min
Non-work time	10 min	11 min
Workplace time	13 h 25 min	17 h 32 min
Logs skidded	46	52
Utilisable volume	108.29 m³	124.02 m³
Volume per log	2.35 m³	2.39 m³

The figures for utilisable volume extracted from both plots differ from that given for the utilisable volume felled, since a number of trees on the study plots were felled after the felling time study had been completed. The felling crew re-entered the plots after skidding had already begun and again searched the area for additional harvestable trees. The additional trees felled were identified on the tree location maps and data on their volumes and the size of the resulting canopy gaps were recorded, although time-study data were not obtained on the felling itself.

Because the skidder was changed partway during the trial study, data on skidding in Plot/01 had to be separated into two sub-samples for data processing; otherwise a comparison between conventional logging and planned harvesting would not have been possible any longer. Consequently, the sample of logs skidded on Plot/02 had also to be divided into two sub-samples for consistency. The sub-samples of 17 logs on Plot/01 and of 20 logs on Plot/02 were spread over approximately 3.3 ha on each plot.

Summary information on the extraction time studies for the sub-samples on the two plots is provided below.

Item	Conventional - Plot/01 Sub-sample (without tree location map)	Planned - Plot/02 Sub-sample (with tree location map)
Work time	6 h 46 min	5 h 57 min
Non-work time	9 min	5 min
Workplace time	6 h 55 min	6 h 02 min
Logs skidded	17	20
Utilisable volume	40.83 m³	49.19 m³
Volume per log	2.40 m³	2.46 m³

Results of the study on extraction for the two sub-samples suggest a clear advantage for planned harvesting over conventional logging in terms of production rates. On average the skidding time per log was 24.38 minutes in conventional logging for the sub-sample on Plot/01 as compared to only 18.10 minutes for the sub-sample on Plot/02 where skidding was supported by the use of the tree location map. The total workplace time per log skidded serves as the basis for estimating production rates, as described in Section 6.2.

Comparing the time required to skid a single log in both systems at the level of individual work elements, only work time is taken into consideration. Table 8 shows the average work time required to skid a single log and its distribution among various work elements.

Table 8. Time study results for log skidding under conventional logging and planned harvesting.
Work elements (classification)		Conventional - Plot/01 sub-sample (without tree location map)		Planned - Plot/02 sub-sample (with tree location map)
		[min]	%	[min]	%
Locating log	(CW)	8.35	35.00	5.90	33.10
Attaching cable to log	(MW)	3.11	13.04	1.10	6.15
Winching	(MW)	0.58	2.41	0.64	3.6
Skidding log(s)	(MW)	7.08	29.66	6.31	35.37
Unhooking	(MW)	1.00	4.17	0.94	5.27
Landing operation	(MW)	0.57	2.40	0.76	4.26
Productive work time	(PW)	20.69	86.68	15.65	87.75
Searching for log	(PL)	2.43	10.17	-	-
Evaluating log	(OP)	0.52	2.17	0.05	0.28
Repairing equipment	(RT)	-	-	1.33	7.44
Maintaining equipment	(MT)	0.23	0.98	0.04	0.22
Opening skidtrails	(AW)	-	-	0.44	2.46
Other	(AW)	-	-	0.33	1.85
Supportive work time	(SW)	3.18	13.32	2.18	12.25
Time required to skid a single log		23.87	100.00	17.83	100.00

The work element "searching for log" refers to the time spent by the skidder operator searching for trees that were felled and bucked. This is considered supportive work since it does not directly contribute to the work progress on an individual work object (IUFRO 1995), and is to be distinguished from the work element "locating logs". The later is part of the productive work time and of each individual cutting cycle where the operator moves by skidder to the next log.

A comparison of productive work times for the two skidding operations suggests a much higher production rate for planned harvesting as compared to conventional logging, with 15.65 minutes required on average to skid one log under planned harvesting versus 20.69 minutes under conventional logging. Moreover, when the additional time spent by the skidding operator searching for logs under conventional logging is considered (more than 10% of total workplace time), the difference is much greater. Clearly, the use of stock maps to assist harvesting provides a substantial improvement in skidding time per log.

6.2 Estimating production rates in felling and extraction

Productivity is defined as the rate of product output per unit of time. Thus, production rates can easily be computed from time-study data in combination with measurements on the outputs of production.

6.2.1 Felling

The estimated production rates for felling operations under both conventional logging and planned harvesting are based on the total time used to produce a certain volume of timber, regardless of whether a particular tree has been felled or rejected. Volume per tree harvested has been calculated by multiplying the average cross-sectional area of the stem by the stem length measured at the felling site. The average cross-sectional area is based on diameter measurements, including bark, made at each end of the stem at the felling site. Times used in the calculations are workplace time excluding meal time.

Table 9. Estimated felling production rates for conventional logging and planned harvesting.
Method, Sample unit	Number of observations	Volume per tree felled [m³]	Felling time per tree [min]	Production rate [m³/h]
Conventional logging - Plot/01 (without tree location map)	40	2.14	13.88	9.25
Planned harvesting - Plot/02 (with tree location map)	44	2.32	10.93	12.74
Notes: 1) Volume per tree felled refers to utilisable volume including bark. 2) Production rates are measured in terms of utilisable volume (including bark) per hour of workplace time but excluding meal times, which were highly variable during the study.

The average utilisable volume per tree felled on Plot/01 was somewhat lower for the study sample than that on Plot/02, although the values for the overall average utilisable volume removed from both plots during harvesting operations, 2.35 m³ and 2.39 m³, are almost identical. As previously indicated, time study data were not obtained on trees felled when the chainsaw operator re-entered both plots to cut additional trees after the skidding operation had already commenced.

Based on the production rates in Table 9 and an assumed effective workplace time of 7 hours per day (excluding meal time), the average production rate would amount to about 65 m³ for the felling crew per day under conventional logging, corresponding to about 30 felled trees per crew-day. The equivalent production rate under planned harvesting in this study was 89 m³ per day, corresponding to about 38 felled trees per crew-day.

6.2.2 Extraction

Estimated production rates for extraction operations under conventional logging and planned harvesting are based on the total time used to provide a certain volume of timber. Volume per load was calculated by multiplying the average cross-sectional area of the stem by the stem length. The average cross-sectional area is based on diameter measurements, including bark, made at each end of the log. Times used in the calculations are workplace time excluding meal time.

Table 10. Estimated extraction production rates for conventional logging and planned harvesting.
Method, Sample unit	Number of observations	Volume per load [m3]	Time required per load [min]	Productivity [m³/h]
Conventional logging - Plot/01 (sub-sample) (without tree location map)	17	2.40	24.38	5.91
Planned harvesting - Plot/02 (sub-sample ) (with tree location map)	20	2.46	18.10	8.15
Notes: 1) Volume per load felled refers to utilisable volume including bark. 2) Production rates are measured in terms of utilisable volume (including bark) per hour of workplace time but excluding meal times, which were highly variable during the study.

Based on the hourly productivity in Table 10 and an effective workplace time of 7 hours per day (excluding meal time), the average production rate is about 41 m³ for the skidding crew per day in conventional logging, representing about 17 logs extracted per crew-day. The corresponding figures under planned harvesting are an average skidding production rate of about 57 m³ per day and 23 logs extracted per crew-day.

Since the same crew operated the Franklin skidder on both study plots, the estimated production rates are comparable for the two operations. The skidder operator had about five years of experience in operating a skidder.

6.3 Cost estimates for harvesting operations

Estimated production costs are based on the harvesting production rates and the hourly costs for the harvesting workforce involved in harvesting activities. The hourly cost estimates are taken from information provided in Whiteman (1999a), updated as indicated in Table 6 following discussions in the field with the concessionaire, staff from SBB, the felling crew, and the logging contractor.

To obtain hourly costs for the skidding crew, the monthly costs in Table 6 were divided by 154 working hours, assuming 22 working days per month and 7 working hours per day. Hourly costs for the felling crew were derived by multiplying the cost per cubic metre by the felling productivity rate under conventional logging as measured on Plot/01. The cost per cubic metre was been calculated by dividing the lump sum paid per tree, which the chainsaw operator received from the concessionaire, by the average log volume per tree at the study site.

Table 11 summarises the labour cost per cubic metre of logs delivered to the landing for those harvesting activities that are carried out differently in the two systems under consideration. These include the pre-harvest inventory of principal tree species and the felling and skidding operations. Costs for equipment and other costs such as road infrastructure and long-distance transport are not included. The cost per cubic metre of the pre-harvest inventory has been derived by multiplying the hourly cost of the inventory crew by the time spent on the inventory on Plot/02, and dividing by the log volume harvested from Plot/02.

Table 11. Estimated labour costs for conventional and planned harvesting.
Activities	Conventional logging-Plot/01 (without tree location map)		Planned harvesting-Plot/02 (with tree location map)
	Productivity [m³/h]	Cost [Sf/m³]	Productivity [m³/h]	Cost [Sf/m³]
Pre-harvest inventory	n/a	-	n/a	92.6
Felling	9.25	759.5	12.74	551.4
Skidding	5.91	1112.6	8.15	806.8
Total costs		1872.1		1450.8
Note: n/a = not applicable

If the labour cost of conventional logging per cubic metre of logs delivered to the landing is assigned an index value of 100%, then the corresponding cost index for planned harvesting, as applied on sample unit Plot/02, is 77.5%. Thus the higher productivity in felling and skidding as a result of planned harvesting not only offsets the additional cost of the pre-harvest inventory but also reduces the overall labour cost by more than 20 per cent. Furthermore, it is important to note that this significant reduction in cost was achieved with untrained operators who are not used to working other than in the conventional way.

The introduction of an hourly payment system might be considered advantageous from the concessionaire's point of view, but only if tree mapping and skidtrail planning were implemented as outlined in the study. Resistance to change to a time-based payment system might be overcome by introducing a "mixed" payment system. The hourly base payment would be supplemented with payment of a piece rate after the production rate exceeds a certain threshold. This would provide an incentive for the operators since they could generate a higher income, but at the same time it would ensure a certain minimum level of timber production. The level would be determined by the average production rate that an untrained operator could be expected to achieve.

6.4 Estimated wood waste

Timber losses for the two logging systems studied have been estimated through a wood recovery study in which sources of timber losses in the felling and extraction operations were analysed. For the felling operations, losses caused by splitting of logs and undiscovered decay in the stem were recorded. For skidding, losses recorded included hang-ups that were not recovered and felled trees that were not found by the skidding crew.

The estimated of wood recoveries for both conventional logging and planned harvesting are summarised in Table 12, with all calculations based on diameter measurements including bark. Table 13 summarises timber losses associated with the two harvesting systems.

Table 12. Wood recovery data on the two study plots.
Study area	Log volume harvested [m³]	Trees harvested	Log volume per tree harvested [m³]	Average stem length [m]	Average stump height [m]
Conventional logging - Plot/01	112.5	48	2.3	13.0	0.90
Planned harvesting - Plot/02	134.9	56	2.4	13.0	0.80

Table 13. Estimated timber losses associated with conventional logging and with planned harvesting on the study area.
Description	Conventional logging-Plot/01 (without tree location map)			Planned harvesting-Plot/02 (with tree location map)
	[m3]	trees	[%]	[m3]	trees	[%]
Total volume felled	133.5	57	100.0	152.8	61	100.0
Volume of logs extracted	112.5	48	84.3	134.9	56	88.3
Total losses	21.0	9	15.7	17.9	5	11.7
Losses due to
hang-ups	4.0	2	3.0	4.7	3	3.1
splitting of logs	-			12.7	1	8.3
undiscovered decay	10.5	4	7.8	-
forgotten trees	6.5	3	4.9	0.5	(1)	(0.3)
The one "forgotten" tree in Plot/02 was intentionally left by the skidding crew because it was a small tree that was buried under the crowns of several other trees. Hence the data are shown here in parentheses.

The figures in Table 13 show that the total timber losses ranged between 11.7% and 15.7% in the study area. The importance of proper felling techniques becomes obvious when considering the loss due to hang-ups and logs split during felling.



A loss in timber volume of about 10% because of poor felling practice should convince authorities, logging associations and concessionaires to seriously consider investing in training programmes. Losses at comparable level were found for poor felling in conventional logging in the studies by Winkler (1997) in the Amazon region in Brazil.

As expected, losses caused by undiscovered decay occurred at the study site, as it also did in conventional logging in the above-mentioned study (Winkler 1997), since acoustic explorations with axe or hammer to detect decay are considered unreliable (Panzer 1991).

Another matter that should be of concern to authorities and concessionaires is the loss of about 5% of the total volume due to lost logs. The use of a tree location map during extraction operations can reduce those losses to zero as demonstrated by the study on Plot/02. Only one small tree of 0.5 m³ was deliberately "forgotten" on that plot by the skidding crew, since it was buried under the debris of several tree crowns.

Figures on avoidable waste of commercial timber reported by Gerwing et al., (1996) on conventional logging done in the State of Para, Brazil, were 6.6 m³/ha or 17.4% of the timber volume from felled trees never located by the skidding crew, and 1.7 m³/ha or 4.5% due to poor felling and bucking practices.



6.5 Estimated damage to residual trees

Harvesting assessments are essential for sustainable forest management since they provide feedback about the quality of the harvesting operations. Unless the forest is left in a condition that will permit the attainment of a desired future condition, sustainability cannot be assured (Dykstra & Heinrich 1996).

Since an estimate of the damage to residual trees and regeneration caused by the harvesting operations is of particular interest in the development of sustainable harvesting technologies, emphasis should be paid to a comprehensive post-harvest assessment. It will provide information on short-term changes in stand structure due to harvesting operations, but possibly also on the medium-term future crop and on conditions for establishment of regeneration of the areas harvested. On this study, however, the post-harvest assessment had to be limited to an assessment of damage to residual trees.

In order to obtain information on damage to residual trees of particular commercial interest, namely PCTs of principle tree species (dbh 25-40 cm), harvesting damage to stems and crowns were recorded and evaluated for each PCT on both sample units Plot/01 and Plot/02. Due to the fact that many harvestable principal tree species with dbh > 40 cm were not cut (see Appendix 4), the assessment of damage was extended to all residual principal tree species.



To separate damages according to their cause, three separate damage assessments were undertaken, as follows:

· prior to felling, since stem and crown damages also occur in primary forests when overmature trees break down;

· after felling, in order to allocate stem damages clearly according to their cause;

· after skidding, in order to assess the potential for improvements since skidding is considered the phase where damages most often occurs through negligence, ignorance, or carelessness rather than as a result of unavoidable circumstances.

Damage to stems (bark and cambium wounds) and crowns were estimated and recorded in damage classes, as follows:

Classification		Crown damage	Stem damage
Class I	slightly damaged	up to 1/3 crown volume damaged	some spots where bark is removed
Class II	seriously damaged	more than 1/3 crown volume damaged	major parts of bark removed (up to 1/3 of girth)
Class III	killed or likely to die	only few branches left / no branches left	tree is not expected to survive; stem broken

Table 14. Stand characteristics on the sample plots before and after harvesting.
Study area	Principal trees species (dbh > 40 cm)				Potential crop trees (principal species) (dbh 25 - 39 cm)
	Volume [m³]	%	Trees	%	Volume [m³]	%	Trees	%
Sample unit Plot/01
Prior to harvesting	623.5	100.0	242	100.0	126.7	100.0	156	100.0
Harvested	128.2	20.6	39	16.1	6.4	5.1	9	5.8
Residual	495.3	79.4	203	83.9	120.3	94.9	147	94.2
Sample unit Plot/02
Prior to harvesting	357.7	100.0	139	100.0	57.3	100.0	81	100.0
Harvested	131.6	36.8	42	30.2	7.8	13.6	14	17.3
Residual	226.1	63.2	97	69.8	49.5	86.4	67	82.7

In general, the inventory on felling and skidding damages is a medium-term assessment with focus on PCTs of principal species. The results stated below might nevertheless be indicative for the future commercial value of the area under review.

Detailed information on all classes of stem and crown damage to PCTs and residual principal tree species with dbh >  40 cm can be found in Appendix 4. Table 15, on the other hand, includes only severe damage (i.e., in damage classes II and III as noted above) due either to natural causes or to the harvesting operation.

Table 15. Natural and harvesting damage to residual trees in the principal tree species.
	Conventional logging - Plot/01 (without tree location map)					Planned harvesting - Plot/02 (with tree location map)
	Number of trees by dbh class					Number of trees by dbh class
Description	25-39 cm	> 40 cm	Total	%	25-39 cm	> 40 cm	Total	%
Trees tallied	147	201	348	100.0	67	94	161	100.0
Damaged trees before felling	0	0	0	0.0	0	2	2	1.2
Damaged trees after felling	5	9	14	4.1	1	3	4	2.5
Damaged trees after skidding	7	11	18	5.5	1	3	4	2.5
Trees without severe damage	140	189	330	94.5	66	91	157	97.5

The number of residual PCTs and other principal tree species without severe damage after harvesting operations is quite high on both sample plots. On Plot/02 where planned harvesting was applied, the natural damage prior to felling was about the same level as the damage which occurred during felling operations, totalling about 2.5%. In conventional logging on Plot/01 the damage was slightly higher, about 5.5%, and with additional damage occurring during skidding.

These results suggest that there should be no major concern about logging damage to the principal tree species and, in particular, the future crop of those species in the next harvest entry. However, due to the fact that the damage inventory was limited to the 15 principal tree species, it should be borne in mind that damage to this small number of species, randomly distributed over the area of the sample units, does not provide a comprehensive picture of the impact on the remaining forest stand, in particular with regard to the skidding operations in the conventional way of logging.



Although the post-harvest assessment was carried out in a similar way as the one in the study by Winkler (1997) in the Brazilian Amazon, the results can hardly be compared due to the quite different number of tree species covered by the studies; 65 commercial species in Brazil against only 15 tree species at this study site in Suriname.

Data reported in the literature on logging damage in natural tropical forests (e.g., Verissimo et al., 1992 and Scharai-Rad et al., 1997) are even more difficult to compare with those in Table 15 because of differences in the damage indices used and in the selection criteria for trees to be used in the analyses.

6.6 Qualitative assessment of environmental impacts

Within the scope of the post-harvest assessments, information was gathered on both sample units in order to estimate environmental impacts associated with felling and ground-skidding operations of planned harvesting as compared to conventional logging.

Survey of skidtrails

The skidtrail systems in sample units Plot/01 and Plot/02 were mapped to obtain information on the skidtrail pattern (see Figures 7 and 8). Due to several delays in the field work and related time constraints, only primary skidtrails were mapped. Skidtrail width and degree of soil disturbance were recorded at intervals along each primary skidtrail.

As mentioned in previous sections, the winching distance of the skidder used in planned harvesting on Plot/02 was limited to about 15 m and the skidder had to move backwards along the cableway in order to reach the logs. Those "secondary skidtrails along the cableway" amounted to about 312 m which could have been avoided had an adequate winch cable been available during the study.

The features of each skidtrail system surveyed in the sample units are shown in Table 16.

Table 16. Skidtrail system as measured in sample units Plot/01 and Plot/02.
Classification of skidtrail	Conventional logging - Plot/01 (without tree location map)			Planned harvesting - Plot/02 (with tree location map)
	width	length	area	width	length	area
	[m]	[m]	[m2]	[m]	[m]	[m2]
Primary	5.82	2.092	12,175	3.93	1,241	4,877
Turnings			260			-
Secondary	n/a	n/a	n/a	(4.08)	(312)	(1,273)

In planned harvesting with designated skidtrails marked in the field before skidding commenced, the area used for primary skidtrails amounts to 5.4% of the 9.1 ha harvested on Plot/02, whereas in conventional logging the corresponding figure is 12.4% of the 10 ha harvested on Plot/01, including turnings. For comparison, Winkler (1997) reported that 4.2% of the area harvested was in primary skidtrails on an environmentally sound harvesting trial in the Brazilian Amazon, as compared to 14.1% in primary skidtrails on a conventional logged area. The conventionally logged area took another 4.6% for secondary skidtrails (Winkler 1997).

Moura-Costa (1997) reports 30-40% of the harvested area traversed by bulldozers in conventional logging operations in tropical rain forests in Sabah, Malaysia.

The average area in roads and landings could not be calculated due a lack of reference areas. As in the vast majority of concessions in Suriname, neither annual coupes nor cutting units have been established in Concession 387. Therefore, the areas actually served by those features of infrastructure are difficult to determine. Figures on this matter reported by Mauro (1996) for tropical rainforests are 3.3% of the logged area in roads and 103 m²/ha in landings.

The importance of a significant reduction in the average area affected by skidtrails per cutting unit as provided by planned forest harvesting, where only 5.4% of the area is used for primary skidtrails, is underscored by the investigations on soil disturbance.



As regards the degree of soil disturbance, about 57% of the total length of primary skidtrails on Plot/01 were attributed to class I. Soil disturbance class I is characterised by the occurrence of only slightly exposed or unexposed mineral soil. Class II, characterised by mineral soil partly or fully exposed and often in combination with gullying, was found for roughly 43% of the total length.

The corresponding figures for class distribution of primary skidtrails on Plot/02 are 49% in class I and about 51% in class II.

Therefore the implementation of environmentally sound harvesting systems, where the skidding machines remain on the designated skidtrails at all times, is deemed highly necessary to permit the attainment of a significant reduction in soil disturbance and soil compaction. Unnecessary damage to residual trees and advance regeneration is also avoided by keeping the skidding machinery on the skidtrails.



6.7 Canopy gaps

Assessments of canopy gaps caused by harvesting operations were carried out in both Plot/01 and Plot/02. These are summarized in Table 17. The width of each gap was along the stem of the felled tree, and further measurements of the gap caused by the tree crown were made. Gaps measured included not only the opening caused by the felled tree itself but also that caused by neighbouring trees that were pulled or knocked down as a consequence of the felling operation.

On average, the size of a canopy gap caused by felling a single tree was 152 mB2 on Plot/01 and about 103 m² on Plot/02. Since felling was done on both plots by the same chainsaw operator, however, this difference appears merely to be the result of chance occurrences.

Average opening of the tree canopy on the two plots was 7.7% and 6.5% for Plot/01 and Plot/02 respectively, a result that is similar to natural canopy opening as a result of tree falls. Uhl & Vieira (1989) report that Amazonian forests normally have 5-20% of their area in gaps at any time. This suggests that at the current level of harvesting intensity, problems associated with large canopy gaps should not be a major consideration in Surinam's forests. Such problems include the germination of vine seeds in response to elevated light and temperature levels (Uhl & Vieira 1989, Putz 1984); the rapid regeneration of pioneer growth such as palms in large canopy gaps (Silva & Whitmore 1990); an increased probability of windfalls when the forest is left in an open, fragmented state; and the threat of fire due to increased solar radiation reaching the forest floor and drying out woody debris in the gaps (Kauffman et al., 1988, Uhl & Vieira 1989, Uhl & Buschbacher 1985).





Table 17. Canopy gaps and canopy opening of the stand due to harvesting.
Study area and harvesting system	Size of study area [ha]	Number of trees	Size of individual canopy gap		Canopy opening [%]
			mean [m2]	maximum [m2]
Plot/01 Conventional logging (without tree location map)	10.0	51	151.54	325.0	7.7
Plot/02 Planned harvesting (with tree location map)	10.0	57	103.11	758.5	6.5 *
* Canopy opening for Plot/02 is based on a reduced sample unit area of 9.1 ha, the area actually harvested within the plot. The remainder of the sample unit was a swampy area of about 9,000 m2. Note: Only trees or parts of trees within each 10 ha sample plot were included in calculating the canopy area affected by harvesting operations.