Stand characteristics, especially the height:diameter ratio (H/D) can greatly effect which leave tree can be effectively used in an area. If the stand is very dense with a high H/D, 100 or higher, it is very likely that trees left in dispersed retention may blow over due to small root wads or break due to rain or snow loads on their crowns. This defeats the purpose of the leave trees if they are doomed to be windthrown or broken. An effective strategy in this case would be to clump the leave trees, especially if they can be sheltered from the wind in a draw or behind a rise. This allows the clump of trees to act like a stand with neighboring trees helping to hold each other up against the forces of the wind. Some trees along the edge of the clump will be to be lost to wind throw.

If the H/D high but the trees have enough of a crown to react to release by thinning clumps of leave trees may be thinned to a lower density than the original stand. This strategy may have many advantages over leaving the stand at the original density. First, by selectively thinning the leave tree area it is possible to leave trees of a variety of sizes and heights available in the leave tree clump. Thinning in this manner will result in a more structurally diverse clump than the original stand, which may attract more wildlife than if the clump was left at its original density. Second, by thinning to a lower density the leave tree clumps will need to cover a larger area than if left at the original density. This area may be either left in one large clump or several smaller clumps. This will provide more area for to be used as wildlife habitat so that more animals may be able to use the clump or will have travel a shorter distance between clumps when dispersing. Third, leaving the trees clumped at a lower density will allow some of the benefits of keeping the leave tree clump dense, support and protection, while increasing the benefits to wildlife. This will also provide a seed source for regeneration while not shading a large portion of the unit.

If the H/D is low enough, less than 80, for example, for the trees to be wind firm they may be left dispersed. This method has the advantages of providing a continuous evenly distributed canopy cover for wildlife and a uniform seed source for regenerating the unit. Leave trees if they are able to respond to the release will grow quickly becoming larger in diameter. Since there are no other trees near to compete for growing space the bottom of the crown will stay at the same height and the branches will increase in diameter. This will provide structure that wildlife species may prefer in the future. One disadvantage to this strategy is the amount of the unit that is in partial shade. This may have an adverse effect on regeneration of Douglas fir. If the unit is to be manage for wood production the decrease in growth of young Douglas-fir trees under the remaining overstory, and hence the harvest volume at rotation, may offset the benefits to wildlife.

The operational feasibility of the leave tree strategy should be considered as well when examining leave tree alternatives. Professor Peter Schiess and Dean Rae Berg did work with regard to loggers' preference to yarding direction, yarding distance, and leave tree retention type in the article Setting Design Evaluation Incorporating Pacific Northwest Loggers' Preference. Loggers' preference for a situation with regards to a combination of a distance factor, and/or a direction factor, and/or silvicultural factors were measured in Utility Values (u). Findings from this are shown in Table 38 and Table 39.

Table 38. Utility values, u, for yarding distance and direction combined (From Schiess and Berg, 1996)

Logging systems were generally designed to yard logs in a certain direction. Cable systems work best yarding up hill whereas ground systems work best yarding down hill. This can be seen in loggers' preference for logging direction with respect to yarding distance in table 1. Downhill yarding is greatly preferred for ground systems over uphill yarding. The opposite is true for cable systems. Loggers prefer uphill yarding to downhill yarding. In the cable case logging is typically done on steep slopes (>30%). When yarding downhill logs can tend to outrun the carriage or butt-rigging and hang up, come loose from their chokers, or knock loose logs laying on the ground waiting to be yarded and send them rolling down the hill towards the machines and crew on the landing. This makes this type of yarding very dangerous to both men and equipment. Loggers prefer downhill yarding with ground systems over uphill since yarding uphill requires more power from the machine to carry the logs against gravity. Thus, it is more efficient to yard downhill with ground based machines. Loggers, regardless of system type, prefer shorter yarding distances. The shorter the yarding distance the shorter the turn time and the higher the daily production will be.

Table 39. Utility values, u, for yarding distance (three levels) and yarding direction (three levels) separately (from Schiess and Berg, 1996)

Attributes	Cable Silvicultural Regime			Ground Silvicultural Regime
Yarding Direction	CC	Aggr	Disp	CC	Aggr	Disp
Uphill	57	68	83	12		27
Sidehill	22	12	11	14		6
Downhill	6	17	23	56		66
Yarding Distance
<500'		52	45		64	66
500' - 1000'		30	29		27	22
>1000'		1	1		1	0

Leave tree retention patterns have a bearing on loggers' preference for different silvicultural regimes. Results of this can be seen in Table 39. It should noted that the rankings do not rank preferences across the board. But, rather, that they rank preference within each silvicultural regime. Thus, Table 39 should not be read that loggers prefer dispersed leave trees to clearcuts with no leave trees when yarding uphill. The opposite is true. Given a choice loggers would most surely prefer that all leave trees be left at the edges of the unit rather than in the middle to get in the way of rigging and yarding!

Given that leave trees need to be in the harvest unit and there are no reasons, silviculturally, that the leave trees cannot be left in any of the aforementioned patterns the question then turns to: how should the leave trees be arranged from an operational standpoint? That greatly depends upon the location of the landing in relationship to the unit and the type of harvest system that will be used to harvest the unit.

Consider first a unit that is to be logged with a ground based system. The choices are between dispersed or clumped (aggregated) retention. For a ground based side it is very easy to work around leave tree clumps. Thus, it can be inferred that the u values for that type of retention would be very similar to those for clearcutting. For both cases loggers prefer downhill yarding. Thus the optimal landing and skid road layout would take logs downhill. If the only option is a large amount of sidehill yarding is necessary loggers would rather have clumps rather than dispersed retention. Even though this may be stretching what the table is saying, in order of preference sidehill rates the lowest of the three yarding directions for dispersed retention whereas it ranks second for clearcut and hence clumped retention. One reason for this may by logs wanting to slid downhill out of the skid road when being yarded sidehill and getting hung up or damaging leave trees with dispersed retention.

The case of cable yarding it gets more complicated. In this case uphill yarding is preferred for all retention patterns. In order of preference downhill ranks higher than side hill with both retention patterns. Sidehill ranks lowest due to the tendency for logs to swing out of the logging corridor and get hung up or damage leave trees. Thus if there are leave trees in the unit sidehill yarding should be minimized.

When leaving trees dispersed throughout the unit operational difficulties may result. Turns being yarded into the landing may swing and get hung up in the leave trees thereby lowering yarding production. Dispersed retention will also require more careful felling of the trees. It will be necessary to fall the trees to lead (away from the yarder) for the trees to be efficiently yarded through the leave trees, especially if whole tree yarding is being done.

Leaving thinned clumps will also require skilled fallers. If clumps are narrow enough it may be possible for the faller to fall the so that the can be yarded directly out o the leave tree clump. In cases where the clumps are more than a tree length wide it may be necessary to fall the trees to lead into the yarding corridor and buck and lime tree before yarding much like in commercial thinning.

When analyzing leave strategies it is important to examine the economic impacts of leave tree density and pattern. If the rules require 5 - 8 trees per acre, on average, to be left in the unit then any more trees left for wildlife or other considerations are trees that are not sent to the mill and thus the revenue for the unit will be reduced. Leaving trees in patterns which reduce yarding production, such as with dispersed retention, or clumps in areas which will get in the way of yarding corridors will increase yarding costs and thus reduce the profit from the sale.

One very common leave tree strategy is to clump the leave trees (Figure 83). This method has many advantages. Much of the unit is clearcut thus allowing efficient logging to be done. Very little of the unit is shaded by the leave trees. Thus there is little effect on regeneration of Douglas fir. If the stand has a high average H/D then this method may be used to help keep the leave trees standing. Leaving the trees in clumps allows neighboring trees to support each other thereby aiding unstable trees. Leaving the clump at the pre-harvest density and structure may not provide much structure for wildlife. These leave tree clumps also do not cover much area. This does not allow very much space in the unit for wildlife. If the pre-harvest density is 250tpa the HCP rules will be met by leaving trees on only 3.2% of the unit.

When silviculturally possible, average H/D not excessively high; leaving trees in thinned clumps may be an option (Figure 84). These clumps may be thinned to a lower density leaving trees from dominant, co-dominant, intermediate and suppressed crown classes. This will allow for more structural diversity for wildlife as well as clumps that cover more area.

Operationally these may be left in such a way that logging may still be done efficiently. Depending on the size of the clumps it may take more skill from the fallers and rigging crew to remove the logs from the clumps without doing excessive damage to the residual stand.

Figure 83. Untreated clump at 250 tpa.

Figure 84. Leave tree clump thinned to 140 tpa with 120 ft² of basal area.

A third method of leave tree retention is dispersed retention. With this pattern leave trees are scattered evenly throughout the unit. This has the advantages of looking good on the landscape (Figure 85), an even canopy cover for dispersing wildlife and short distances between wildlife trees. This retention pattern will not provide dense habitat that some species may need for cover. Silvicultrally the trees left should have relatively low H/D's and large crowns. If the H/D is to high the trees are very likely to be windthrown without any trees near for support. Without large crowns the trees may not be able to respond to the increase in growing space and thus may not grow into stable trees.

Figure 85. Dispersed retention of 8tpa

Leaving 8tpa shades a large part of the unit. Even at this level of shading there may be a reduction in growth of Douglas-fir regen. If the goal is to maximize tree growth this may not be a good option due to possible loss of growth in young trees.

Operationally this option is more difficult to work with if the unit is to be logged with a cable system. Dispersed trees become obstacles to the trees being yarded to the landing. Hang-ups are much more likely which result in lower production rates.